New energy vehicle passageway identification and grading method based on ETC pass data

Abstract

The application discloses a new energy vehicle passage corridor identification and grading method based on ETC passage data, and belongs to the technical field of intelligent traffic and expressway network operation management. The method first performs double-checking and cleaning on the original flow data of the expressway ETC, and matches and extracts effective passage records of new energy vehicles. In combination with the new energy vehicle energy supplement time redundancy, a mileage constraint interval is delimited, path reconstruction is completed through two-stage space-time checking, the comprehensive passage intensity of a network unit is calculated by weighting the passage volume and passage proportion of the new energy vehicles, candidate corridors are generated by aggregating continuous high-intensity units, and finally, corridor grading is completed based on multi-dimensional core indexes. The application deeply adapts to the expressway travel characteristics of new energy vehicles, improves the path restoration precision and corridor identification accuracy, and can provide accurate support for the planning of new energy supplement facilities on expressways, and the fine management and control of road networks.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation and highway network operation and management technology, and in particular to a method for identifying and classifying the passage corridors of new energy vehicles based on ETC traffic data. Background Technology

[0002] With the rapid development of my country's new energy vehicle industry, the number of new energy vehicles has continued to grow rapidly, and their share of traffic on the highway network has also increased significantly year by year. Unlike traditional fuel vehicles, new energy vehicles have the core characteristics of limited driving range and rigid demand for high-speed refueling. Refueling directly leads to increased vehicle travel time and deviations from the theoretical shortest path. The spatiotemporal distribution and path selection logic of their high-speed travel are fundamentally different from those of fuel vehicles. Correspondingly, the specific needs of new energy vehicles for high-speed travel, such as refueling services, traffic safety, and travel guidance, are becoming increasingly prominent. This has raised new, precise, scenario-based, and customized business requirements for the planning and layout of new energy refueling facilities, differentiated operation and management, and emergency support system construction of the highway network.

[0003] The ETC (Electronic Toll Collection) system for highways has achieved full coverage of the national road network, stably collecting entrance-exit (OD) traffic flow data for all passing vehicles. It is a core and authoritative data source for understanding the spatiotemporal traffic patterns of vehicles on highways. Vehicle corridor identification based on ETC traffic data has become a mainstream core technology in highway network operation management and transportation infrastructure planning. Existing related technologies have formed a standard technical framework covering data preprocessing, route reconstruction, traffic characteristic statistics, and corridor identification, and are widely used for analyzing road network traffic patterns and supporting operational decisions for all types of vehicles.

[0004] However, existing corridor identification technologies all focus on analyzing all types of vehicles or traditional fuel vehicles, failing to adapt and optimize the underlying technical logic for the unique travel characteristics of new energy vehicles. This makes it impossible to meet the business needs of identifying and classifying dedicated corridors for new energy vehicles. In the path reconstruction stage, existing technologies only use a single level of shortest path constraints and average speed verification, neglecting to consider the dwell time and non-shortest path travel scenarios caused by new energy vehicles' refueling needs. This either results in the rejection of long-duration, long-mileage valid paths caused by refueling as abnormal records, or an inability to accurately reconstruct the actual travel path of vehicles, leading to severely insufficient path reconstruction accuracy. Furthermore, the granularity of existing analysis units is coarse, not refined to the smallest road network unit between interconnections, making it difficult to support the refined identification of continuous corridors. In the traffic feature quantification stage, existing technologies only use a single indicator of absolute traffic volume, failing to consider the proportion of new energy vehicles in the road network and their time-based traffic characteristics. This makes it impossible to distinguish between road sections with high traffic volume for all vehicle types but low proportion of new energy vehicles and road sections with high dedicated traffic for new energy vehicles, failing to accurately represent the high-frequency traffic needs of new energy vehicles. Consequently, the subsequent corridor identification results deviate significantly from the actual travel patterns of new energy vehicles. In the corridor identification process, existing technologies mostly rely on human experience to set screening thresholds, lack standardized topology aggregation rules, and do not consider the layout requirements of charging facilities for long-distance continuous passage of new energy vehicles. The generated passage corridor space is fragmented and has poor continuity, which cannot meet the core business requirement of systematically deploying new energy charging facilities along continuous passage corridors. Summary of the Invention

[0005] This invention overcomes the shortcomings of the prior art and provides a method for identifying and classifying the passage corridors of new energy vehicles based on ETC passage data.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a method for identifying and classifying the passage corridors of new energy vehicles based on ETC (Electronic Toll Collection) data, comprising the following steps:

[0007] S1. Obtain the original ETC transaction data of the highway containing vehicle type code, entrance and exit toll station and corresponding passage time fields. After double verification of time logic and field integrity, abnormal and invalid records are removed. Valid new energy vehicle passage records are extracted based on the new energy vehicle type code lookup table.

[0008] S2. Obtain highway network topology data containing interconnected nodes and basic network unit attributes between adjacent interconnections, and construct a directed adjacency matrix of the road network; take the OD corresponding to the interconnection of a single passage record as the start and end point, and combine the refueling time redundancy of new energy vehicles to delineate the candidate path mileage constraint interval, generate acyclic candidate paths that meet the interval requirements, screen the optimal driving path through two levels of spatiotemporal speed and duration verification, decompose it into an ordered basic network unit sequence, and summarize to generate a full set of unit passage sequences;

[0009] S3. Statistically calculate the traffic volume and traffic ratio of new energy vehicles in each network unit according to the same time window, calculate the time-sharing traffic intensity by weighting, take the average value of the whole time period to generate the comprehensive traffic intensity of each unit, and construct a network unit map with intensity labeling.

[0010] S4. Based on the quantile of the comprehensive traffic intensity of all units, set the traffic intensity threshold, filter high-intensity network units, aggregate continuously connected high-intensity units according to the road network topology connectivity, and generate a candidate traffic corridor set.

[0011] S5. Calculate the three core evaluation indicators of continuity, traffic intensity and time period stability for each candidate corridor to construct an evaluation matrix. After dimensionless processing and objective weighting, calculate the comprehensive evaluation score, complete the level division according to the preset level threshold, and output the leveled list of new energy vehicle traffic corridors.

[0012] In a preferred embodiment of the present invention, in S1, the time logic verification is to remove records where the exit passage time is earlier than the entrance passage time and the passage duration exceeds the reasonable passage interval of OD; the reasonable passage interval of OD is calculated based on the shortest path mileage of OD and the statutory speed limit of the highway; the new energy vehicle type code lookup table is constructed according to the toll road vehicle classification standard of the Ministry of Transport and the unified coding rules of new energy vehicles of each province's network toll collection system.

[0013] In a preferred embodiment of the present invention, in S2, the lower limit of the mileage constraint interval is the mileage of the theoretical shortest path from OD to S2, and the upper limit is the product of the highest legal driving speed and the actual travel time; the refueling time redundancy is reflected by the maximum time redundancy coefficient, which has a value range of 1.2 to 1.8, and is used to cover the time increment of refueling stops and slow driving of new energy vehicles.

[0014] In a preferred embodiment of the present invention, in S2, the first-level verification of the two-level spatiotemporal verification is that the theoretical average driving speed of the candidate path falls within the legal speed range of the OD pair, and the second-level verification is that the longest reasonable driving time of the candidate path's theoretical shortest driving time and the superimposed reasonable refueling stop time covers the actual travel time; the reasonable refueling stop time is the maximum value of a fixed 2 hours and a stop time of 15 minutes per 100 kilometers.

[0015] In a preferred embodiment of the present invention, in S3, the formula for calculating the time-sharing traffic intensity is:

[0016] ;

[0017] In the formula, For the first The network unit in the first Time-based traffic intensity for each time window This represents the traffic volume of new energy vehicles after the min-max standardization. The proportion of new energy vehicles on the road. This is the weighting coefficient for the traffic volume of new energy vehicles. This is the weighting coefficient for the proportion of new energy vehicles, and , The value ranges from 0.5 to 0.7. The value ranges from 0.3 to 0.5.

[0018] In a preferred embodiment of the present invention, in S4, the quantile is taken as the 80th percentile value; the aggregation operation is a recursive aggregation, the rule of which is: traverse the high-intensity units that have not been aggregated, take the unit as the initial node, recursively traverse all directly connected high-intensity units based on the directed adjacency matrix until there are no new units, and combine the continuously connected units into a single candidate passageway.

[0019] In a preferred embodiment of the present invention, in S5:

[0020] The continuity index is calculated based on the number of network units contained in the corridor and the connectivity integrity of the road segment to which it belongs;

[0021] The traffic intensity index is calculated based on the average and sum of the comprehensive traffic intensity of all network units within the corridor;

[0022] The time-period stability index is calculated based on the coefficient of variation of the time-based traffic intensity of all network units within the corridor.

[0023] In a preferred embodiment of the present invention, in step S5, the objective weighting and comprehensive evaluation are completed using the entropy-weighted TOPSIS comprehensive evaluation model as follows:

[0024] S501. The evaluation matrix is ​​subjected to dimensionless processing based on the min-max positive standardization method to obtain the standardized matrix;

[0025] S502. Calculate the objective weights of each evaluation index based on the entropy weight method, and generate a weighted standardized matrix by combining the standardized matrix.

[0026] S503. Determine the positive and negative ideal solutions for each indicator, calculate the Euclidean distance between each candidate corridor and the positive and negative ideal solutions, and calculate the comprehensive proximity score based on the Euclidean distance.

[0027] S504. Based on the preset three-level proximity threshold, complete the level division and output the level list.

[0028] In a preferred embodiment of the present invention, the three-level proximity thresholds are a first threshold, a second threshold, and a third threshold that decrease sequentially; those with a comprehensive proximity greater than or equal to the first threshold are classified as core backbone corridors, those with a first threshold greater than or equal to the comprehensive proximity greater than or equal to the second threshold are classified as important support corridors, those with a second threshold greater than or equal to the comprehensive proximity greater than or equal to the third threshold are classified as potential cultivation corridors, and those with a proximity lower than the third threshold are removed from the list.

[0029] In a preferred embodiment of the present invention, the method further includes a periodic dynamic update mechanism: a fixed update cycle is set, and when the cycle arrives, new ETC transaction data is obtained, and steps S1 to S5 are repeated to generate an updated classification list; the update result is compared and verified with the historical result of the previous cycle, and if the corridor level changes by more than one level, a new core backbone corridor is added, or an existing core backbone corridor is removed, a level change warning message is output.

[0030] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0031] (1) For the specific scenario of increased dwell time and deviation of driving path from the shortest path caused by the demand for energy replenishment of new energy vehicles, a progressive path reconstruction system was constructed, which consists of mileage constraint interval delineation, acyclic candidate path generation, and two-level spatiotemporal speed and duration verification. Through the special design of energy replenishment time redundancy coefficient and reasonable dwell time for energy replenishment, the core defects of existing technology in misjudging effective energy replenishment paths and logical contradictions in path restoration were solved. At the same time, with the basic network unit of highways between interchanges as the smallest analysis granularity, the accurate mapping of ETC toll OD records to the smallest unit of the road network was realized, laying a solid spatial foundation for the refined identification of continuous passage corridors.

[0032] (2) To address the characteristic bias caused by the use of only a single indicator of absolute traffic volume in existing technologies, a time-sharing and comprehensive dual-dimensional quantitative system for the exclusive traffic intensity of new energy vehicles was constructed. This system integrates the standardized traffic volume of new energy vehicles and the proportion of traffic in the road network as dual core indicators. It not only covers the overall traffic characteristics of all time periods, but also highlights the exclusive traffic attributes of new energy vehicles. It can accurately distinguish between road network units with large traffic volume of all vehicle types but low proportion of new energy vehicles and those with exclusive high-frequency traffic of new energy vehicles. This provides an exclusive and objective quantitative basis for the screening of high-intensity traffic units and the identification of corridors.

[0033] (3) In response to the problem that existing technologies rely on human experience to set thresholds and generate fragmented corridors, the quantile rule is used to adaptively set the traffic intensity threshold, avoiding the subjective bias brought by human experience. It can adapt to the differences in road network characteristics in different regions and time periods. At the same time, a standardized recursive aggregation rule is established based on the directed adjacency matrix of the road network, realizing the automated aggregation of continuous high-intensity units. The generated traffic corridor has complete spatial continuity, perfectly meeting the core business needs of the systematic layout of new energy replenishment facilities along continuous traffic corridors.

[0034] (4) In response to the problem of subjective weighting and serious bias in the classification results of existing technologies, three core evaluation indicators that fit the needs of long-distance continuous travel and energy replenishment facility layout of new energy vehicles are selected: continuity, traffic intensity and time period stability. An objective weighting method based on the discrete characteristics of data is adopted to eliminate the result bias caused by subjective weighting. At the same time, the entropy weight-TOPSIS model is used as the preferred implementation method, focusing the core invention on the path reconstruction and continuous corridor identification mechanism for new energy vehicles. This not only ensures the objectivity and accuracy of the classification results, but also improves the stability of patent authorization. The classification results can provide accurate data support for the differentiated energy replenishment planning and operation management strategy formulation of different levels of traffic corridors.

[0035] (5) This invention takes the energy replenishment demand, range constraints, and route selection logic of new energy vehicles as core constraints, and runs through the entire technical process of data preprocessing, route reconstruction, feature quantification, corridor identification, and hierarchical evaluation. It realizes the upgrade from application object adaptation to the underlying technical logic adaptation, and solves the core pain point of insufficient adaptability of existing technologies to the travel characteristics of new energy vehicles. At the same time, it is equipped with a periodic dynamic update and level change early warning mechanism, which can adapt to the dynamic changes of the high-speed travel pattern of new energy vehicles in real time, and provide long-term, stable and accurate decision support for the planning and refined operation and management of new energy replenishment facilities in the highway network. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart of a preferred embodiment of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0040] like Figure 1 As shown, the method for identifying and classifying the passage corridors of new energy vehicles based on ETC (Electronic Toll Collection) data includes the following steps:

[0041] S1. Obtain the original ETC transaction data of the highway, which includes the vehicle type code, entrance and exit toll station and corresponding passage time fields. After double verification of time logic and field integrity, abnormal and invalid records are removed. Valid new energy vehicle passage records are extracted based on the new energy vehicle type code lookup table.

[0042] Furthermore, the process of generating valid new energy vehicle passage records includes:

[0043] S101. The original ETC passage data obtained includes the fields of vehicle license plate number, vehicle type code, entrance toll station number, entrance passage time, exit toll station number, and exit passage time.

[0044] S102. First, based on the time verification rules, remove abnormal records where the exit passage time is earlier than the entrance passage time or the passage duration exceeds the preset reasonable threshold. Then, based on the field integrity rules, remove invalid records where the core fields are empty.

[0045] S103. Based on the new energy vehicle type code lookup table, match the vehicle type codes in the cleaned valid records, identify and extract the passage records corresponding to all new energy vehicles, and generate valid new energy vehicle passage records.

[0046] In practical implementation, the basic data input of this method is divided into two categories. The first category is the full amount of original ETC passage data output by the provincial expressway network toll collection center. The original data includes the core fields of vehicle license plate (including license plate color), vehicle type code, entrance toll station number, entrance toll station geographical coordinates, entrance passage time, exit toll station number, exit toll station geographical coordinates, exit passage time, passage medium number, and transaction serial number.

[0047] The second category is the vector topology data of the expressway network released by the provincial transportation authorities. It includes three core elements: toll station node set, interchange node set, and basic highway network unit set. The basic highway network unit is defined as a continuous and indivisible road segment between two adjacent interchange nodes in the expressway network. It is the smallest unit for the network analysis in this method. Each basic network unit is configured with a unique unit code, starting and ending interchange node number, road segment number, mileage attribute, and design speed limit attribute.

[0048] In the data preprocessing and accurate identification of new energy vehicles, the original ETC toll data is first cleaned using rules to remove invalid and abnormal records. For each individual original toll record, a time-based logical check is performed, defining the toll duration for each record as... The calculation formula is:

[0049] ;

[0050] In the formula, This represents the exit passage time for this record, in hours. This represents the entry passage time for this record, in hours. Simultaneously, based on the OD pairs corresponding to the entry and exit toll stations for this record, the shortest path mileage of the road network is calculated. The unit is km. Considering the maximum speed limit of 120 km / h and the minimum speed limit of 60 km / h on highways, the reasonable travel time range for this OD pair is determined using the following formula:

[0051] ;

[0052] ;

[0053] In the formula, This represents the theoretical shortest travel time for the OD pair, in hours. This represents the theoretically longest reasonable travel time for this OD pair, in hours. After calculation, discard... , , After checking for time-related logical errors and abnormal records, a field integrity check is performed to remove records where any core field such as vehicle license plate number, vehicle type code, entrance toll station number, entrance passage time, exit toll station number, or exit passage time is empty or has an invalid code. The final result is a cleaned set of valid passage records for all vehicle types, denoted as […]. A single record is denoted as , , This represents the total number of valid passage records for all vehicle types.

[0054] After data cleaning, precise identification of new energy vehicles is performed. Based on the Ministry of Transport's "Vehicle Classification for Toll Roads" (JT / T 4892019) and the unified coding rules for new energy vehicles of various provincial network toll centers, a set of new energy vehicle type codes is constructed. ;in, For the first Type codes for compliant new energy vehicles. Traverse the valid set of all vehicle type access records. Each record in Extract its vehicle type code ,like Then the record will be included in the set of valid new energy vehicle passage records. The set expression is:

[0055] ;

[0056] In the formula, The final output set of valid new energy vehicle passage records is denoted as follows: , , This represents the total number of valid new energy vehicle passage records.

[0057] S2. Obtain highway network topology data containing interconnected nodes and basic network unit attributes between adjacent interconnections, and construct a directed adjacency matrix of the road network; take the OD corresponding to the interconnection of a single passage record as the start and end point, and combine the refueling time redundancy of new energy vehicles to delineate the candidate path mileage constraint interval, generate acyclic candidate paths that meet the interval requirements, screen the optimal driving path through two-level spatiotemporal speed and duration verification, decompose it into an ordered basic network unit sequence, and summarize to generate a full set of unit passage sequences.

[0058] Furthermore, the process of generating the complete network unit traffic sequence set includes:

[0059] S201. Obtain highway network topology data including the geographic coordinates of each toll station, the start and end nodes of each basic highway network unit, connection relationships, and mileage attributes.

[0060] S202. Taking the entrance toll station of a single valid new energy vehicle passage record as the starting point and the exit toll station as the ending point, calculate the reasonable driving speed range by combining the passage time difference between the entrance and the exit, perform path traversal based on the connectivity of the road network topology, and select the optimal driving path that meets the driving speed range constraint.

[0061] S203. Decompose the optimal driving path into ordered connected basic road network units, generate network unit travel sequences corresponding to single travel records, and generate a full set of network unit travel sequences after summarizing all records.

[0062] The basic network unit sequence of highways is the road segment unit between two adjacent interchange nodes in the expressway network. Each basic network unit of highways is configured with a unique unit code, the corresponding start and end interchange node number, the road segment number, and the mileage attribute.

[0063] During the path reconstruction operation, the uniqueness of the unit code is checked for each restored path to ensure that the unit codes in a single network unit travel sequence are arranged in order according to the travel direction and without duplicate codes, and invalid travel sequences with duplicate unit codes or logical errors in order are eliminated.

[0064] In practical implementation, during the refined reconstruction of traffic routes, the first step is to construct a road network connectivity adjacency matrix A based on highway network topology data, with a matrix dimension of [missing information]. , The total number of basic network units within the road network, matrix elements The rule for determining the value is: if the basic network unit Endpoint interconnection nodes and basic network units If the starting interchange nodes are the same node and the driving directions are the same, then , indicates that the two units are directly connected; otherwise , represents no direct connection relationship, where, For the first in the road network A basic network unit, .

[0065] Targeting the valid traffic records of new energy vehicles Each single record in Starting from the access and interconnection node corresponding to its entrance toll station. The access and interconnection node corresponding to the exit toll station is the endpoint. Combined with the entry passage time of this record Export passage time Calculate the actual travel time The unit is hours. The reasonable driving speed range for this record is calculated simultaneously using the following formula:

[0066] ;

[0067] ;

[0068] In the formula, This is the total mileage of the shortest path in the road network for the corresponding OD pair, in km; 0.8 and 1.2 are speed fluctuation coefficients used to cover the mileage deviation between actual driving and the shortest path, avoiding overly strict path selection; The minimum reasonable driving speed corresponding to this record, in km / h; The maximum reasonable driving speed is expressed in km / h.

[0069] Based on the constructed road network connectivity adjacency matrix A, with Starting point Using Dijkstra's shortest path algorithm as the destination, a full path traversal is performed to obtain a set of all candidate paths connecting the starting point and the destination. Each candidate path corresponds to a total mileage. Composed of an ordered sequence of basic network units, the theoretical average speed of the candidate path is calculated. ,like If the candidate path is found to be the optimal driving path that meets the constraints, then the candidate path is determined to be the optimal driving path.

[0070] The basic network unit sequence obtained from the optimal travel path decomposition is subjected to dual verification: first, a sequential logic verification, where the endpoint interconnected node of the previous unit in the sequence must be the starting interconnected node of the next unit, conforming to the connectivity rules of the adjacency matrix A; second, a unit code uniqueness verification, where no duplicate unit codes appear in the sequence to avoid path loop backtracking errors. If the verification fails, the path traversal is re-executed; if the verification passes, a valid network unit travel sequence corresponding to that travel record is generated. ;in, This is the first basic network unit that connects to the network. The last basic network unit accessed at the endpoint. This represents the number of basic network units contained in the sequence. Finally, all valid passage records' corresponding network unit passage sequences are summarized to generate a complete set of network unit passage sequences. .

[0071] It should be further explained that step S2 addresses the technical deficiency of ETC toll records, which only contain OD start and end information and lack intermediate travel trajectories. It achieves accurate convergence of the feasible solution space through shortest path constraints and accurately restores the actual travel path of the vehicle through multi-dimensional spatiotemporal matching rules. Finally, it maps a single OD toll record into an ordered and traceable sequence of basic highway network units, providing accurate spatial granular data support for subsequent traffic intensity quantification and traffic corridor identification.

[0072] This method employs a standardized directed weighted graph model from the field of graph theory to digitally model the highway network. The model expression is as follows: ,in:

[0073] Node set It is a collection of all interconnected hub nodes, toll station access nodes, and road segment boundary nodes within the road network. Each node is configured with three core attributes: a unique ID, WGS84 geodetic coordinates, and the road segment number to which it belongs. It serves as a spatial anchor point for path search.

[0074] Edge set The basic network unit, as defined in the claims, is a continuous one-way road segment between two adjacent nodes and is the smallest unit for road network analysis in this method. Each basic network unit is configured with a unique unit code, start and end node IDs, and one-way mileage. (Unit: km) Design maximum speed limit (Unit: km / h), driving direction, and the highway number are six fixed attributes.

[0075] Weight set , where is the mileage weight matrix corresponding to the edge set, where Indicates from node To the node The unidirectional mileage of the basic network unit, when there are no directly connected edges. .

[0076] Based on the above directed weighted graph, construct the directed adjacency matrix of the road network. The matrix dimension is ( (This represents the total number of basic network units), and the rules for selecting the values ​​of the matrix elements are as follows:

[0077] .

[0078] This matrix provides a standardized basis for topology judgment in subsequent path traversal and connectivity verification.

[0079] The output of step S1 in this section is a set of valid new energy vehicle passage records, with each valid record being a single record. The following standardized fields must be included: entrance toll station number, and the access node corresponding to the entrance toll station. Entrance passage time Exit toll station number, corresponding access node of the exit toll station Export passage time All time fields are uniformly converted to hour-level Unix timestamps.

[0080] For a single valid record, the following core basic parameters are pre-calculated:

[0081] Actual total driving time The calculation formula is:

[0082] ;

[0083] In the formula, The unit is hourly; the preceding data cleaning process has been eliminated. , Recording abnormal entries that exceed the reasonable threshold for OD ensures the validity of parameters.

[0084] OD to theoretical shortest path mileage :by Starting point As the endpoint, using the mileage weight matrix The values ​​were calculated using Dijkstra's shortest path algorithm, with units of km, and are the core benchmark values ​​for the shortest path constraint.

[0085] The core function of the shortest path constraint is to define the boundary of the feasible solution space for path search, rather than forcing vehicles to choose the shortest path. By quantifying the mileage constraint interval, it eliminates invalid paths that are too long, detour, or do not conform to travel logic from the root, while significantly reducing the computational complexity of traversing the entire road network.

[0086] Based on OD (Original Design) to find the shortest path mileage in theory Compared with actual driving time Based on the maximum speed limit of 120 km / h and the minimum speed limit of 60 km / h for expressways stipulated in the "Road Traffic Safety Law of the People's Republic of China", the mileage constraint interval of the candidate path is quantitatively defined as follows:

[0087] ;

[0088] ;

[0089] ;

[0090] In the formula:

[0091] The total mileage of a single candidate path, in km; The maximum allowed mileage of the candidate path is the upper boundary of the feasible solution space; 0.7 represents the maximum legal driving speed corresponding to the OD pair, in km / h; 0.7 is the minimum time compression factor, used to cover scenarios where the actual path is slightly longer than the shortest path, and can be dynamically adjusted according to the characteristics of the regional road network, with a value range of [0.5, 0.8].

[0092] The core logic of the spatiotemporal matching rule is to verify the spatiotemporal consistency of path mileage, travel time, and road speed limits. Through multi-level progressive quantitative verification, it selects the optimal path with the highest matching degree to the actual driving behavior from the candidate path set. It includes two levels of core verification rules:

[0093] For a single candidate path Calculate its theoretical average driving speed To verify whether it falls within the legal speed range, the calculation formula and verification rules are as follows:

[0094] ;

[0095] ;

[0096] ;

[0097] In the formula, 1.3 represents the minimum legal driving speed corresponding to the OD pair, in km / h; 1.3 is the maximum time redundancy coefficient, used to cover scenarios such as vehicle stops at service areas, slow-moving traffic, and refueling of new energy vehicles. It can be dynamically adjusted according to the characteristics of the regional road network, and its value range is [value missing]. .

[0098] For candidate routes that pass the first-level verification, further verification is conducted to confirm the matching between their segmented travel time and the speed limits of each road segment, resolving the technical defect of compliant average speed but contradictory segmented travel logic. The ordered sequence of basic network units after decomposition is as follows: Calculate its theoretical shortest travel time Compared with the theoretical longest reasonable driving time The verification rules are as follows:

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] In the formula: The shortest theoretical time, in hours, for a vehicle to travel along the entire route at the maximum speed limit designed for each section without any stops; The longest reasonable duration, in hours, is the total time a vehicle can travel along the entire route at the legal minimum speed limit, plus reasonable stop times. To determine the maximum reasonable dwell time, in hours, we take the maximum value of a fixed 2 hours and a dwell time of 15 minutes per 100 kilometers, which aligns with the refueling and rest characteristics of new energy vehicles traveling at high speeds. , They are the first in the sequence The mileage and design maximum speed limit of each basic network unit.

[0104] Shortest path constraints and spatiotemporal matching rules form a progressive collaborative computing framework:

[0105] Phase 1: By using the shortest path constraint, the infinite path search space of the entire road network is converged to a finite mileage constraint interval, generating only candidate paths that meet the interval requirements.

[0106] The second stage involves performing multi-level consistency checks on the converged candidate path set using spatiotemporal matching rules to eliminate paths with spatiotemporal logical contradictions, and finally locking in the optimal path with the highest degree of matching with the actual driving behavior of the vehicle.

[0107] For a single valid new energy vehicle passage record, perform the following calculations:

[0108] Mapping OD to the corresponding starting access node Access node to the endpoint Dijkstra's algorithm is used to calculate the OD pair of the theoretical shortest path mileage. ;

[0109] Calculate the total actual driving time And calculate the maximum legal driving speed. Minimum legal driving speed Maximum permissible mileage Finally, the mileage constraint interval of the candidate path is determined. .

[0110] For the defined feasible solution space, the Yen's K Shortest Paths algorithm is used to generate a set of acyclic candidate paths. The specific execution rules are as follows:

[0111] by Starting point As the endpoint, using the mileage weight matrix As the basis for calculation, set The value is 10 (can be adjusted to 15 in densely populated hub areas), before generation. A simple acyclic path with increasing mileage.

[0112] Calculate the total mileage for each generated path. Only retain those that meet the requirements. The path, based on the directed adjacency matrix. Invalid paths that contain reverse paths, loops, or disconnected nodes are eliminated, ultimately forming a candidate path set. ;

[0113] Each candidate path is decomposed into an ordered sequence of basic network units, where adjacent units in the sequence satisfy... The connectivity rules provide standardized input for subsequent spatiotemporal matching.

[0114] For the candidate path set, a three-level progressive matching and filtering process is performed to determine the unique optimal path:

[0115] Calculate the theoretical average driving speed for each candidate path. Eliminate those that do not meet the requirements. The path is followed to complete the initial screening;

[0116] For the paths that pass the initial screening, calculate and Eliminate those that do not meet the requirements. The path is used to complete fine-grained verification;

[0117] For paths that pass the two-level verification, calculate the spatiotemporal matching degree. The path with the highest matching degree is selected as the final optimal driving path. The matching degree is calculated using the following formula:

[0118] ;

[0119] In the formula, The default value is 90 km / h, which represents the historical average driving speed of the OD for the local road network. It can be optimized iteratively through historical ETC big data. The range of values ​​is The closer the value is to 1, the higher the matching degree.

[0120] If two matches are found with the same mileage, the match with the closest mileage will be selected first. The route is designed to align with the typical behavioral characteristics of highway travelers.

[0121] For the determined optimal driving path, perform double validity checks to ensure the sequence meets the requirements of subsequent analysis:

[0122] The verification sequence shows that the endpoint node of the preceding basic network unit is exactly the same as the starting node of the following unit, which conforms to the directed adjacency matrix. The connectivity rules are such that there are no breakpoints and no reverse paths.

[0123] The basic network unit codes that are not repeated in the verification sequence are excluded, and invalid sequences such as loops and reversals that do not conform to the normal driving logic of high speed are excluded.

[0124] If the verification passes, a valid network unit passage sequence corresponding to the passage record is generated; if the verification fails, the candidate path set is returned, and the path with the second highest matching degree is selected for re-verification; if all paths fail the verification, the record is marked as abnormal and removed from the dataset. Finally, the network unit passage sequences corresponding to all valid new energy vehicle passage records are summarized.

[0125] S3. Statistically calculate the traffic volume and traffic ratio of new energy vehicles in each network unit according to the same time window, calculate the time-sharing traffic intensity by weighting, take the average value of the whole time period to generate the comprehensive traffic intensity of each unit, and construct a network unit map with intensity labeling.

[0126] Furthermore, the process of generating the network unit map includes:

[0127] S301. Divide a single day of 24 hours into several windows of equal length. Based on the full network unit traffic sequence set, calculate the traffic volume of new energy vehicles and the total traffic volume of all types of vehicles in each network unit within each time window.

[0128] S302. Calculate the proportion of new energy vehicles in the corresponding time window of each network unit, and perform a weighted summation operation on the traffic volume of new energy vehicles and the proportion of new energy vehicles to obtain the time-sharing traffic intensity of each network unit in the corresponding time window.

[0129] S303. Perform average calculation on the time-sharing traffic intensity of the entire day's time window to generate the comprehensive traffic intensity of each network unit, that is, generate a network unit map with traffic intensity feature annotations.

[0130] In practical implementation, during the network unit traffic strength quantification and characterization stage, a single day of 24 hours is first divided into T equal-length time windows. In this embodiment, 1 hour is used as the time window length, i.e., T=24. Each time window is denoted as... , The corresponding day's Hours.

[0131] Based on the full network unit traffic sequence set Statistics on each basic network unit In each time window The two core traffic flow indicators are: firstly, the traffic volume of new energy vehicles. That is, all containing In the passage sequence, the passage time covers the time window. The total number of new energy vehicles; and the total traffic volume of all types of vehicles. That is, the corresponding time window Inside, all vehicle types' valid passage records are collected and processed. The total number of vehicles.

[0132] Based on the statistical results, the calculation of each basic network unit is performed. In the time window The proportion of new energy vehicles in the country The calculation formula is:

[0133] ;

[0134] In the formula, if If no vehicles pass through during that period, then , The range of values ​​is .

[0135] To eliminate the difference in dimensions, the traffic volume of new energy vehicles is standardized using the minmax method to obtain the standardized traffic volume. The calculation formula is:

[0136] ;

[0137] In the formula, This represents the minimum traffic volume of new energy vehicles across all network units at all times. This represents the maximum number of new energy vehicles passing through all network units at all times. The range of values ​​is .

[0138] Based on the standardized traffic volume and the proportion of new energy vehicles, the calculation of each basic network unit is performed. In the time window Time-sharing traffic intensity within The calculation formula is:

[0139] ;

[0140] In the formula, This is the weighting coefficient for the traffic volume of new energy vehicles. In this embodiment, the weighting coefficient for the proportion of new energy vehicles is... , The weighting coefficients can be adjusted according to operational needs and meet the requirements. ; The time-sharing traffic intensity has a value range of [value range missing]. .

[0141] Calculate each basic network unit based on time-division traffic intensity. Overall traffic intensity The calculation formula is:

[0142] ;

[0143] In the formula, In this embodiment, the total number of time windows divided within a single day is [not specified]. , The arithmetic mean of the time-sharing traffic intensity across the entire daily time window represents the overall traffic activity of new energy vehicles in this network unit. Finally, the geospatial information and topological relationships of all basic network units are correlated and mapped with their corresponding comprehensive traffic intensity to generate a network unit map labeled with traffic intensity features.

[0144] S4. Based on the quantile of the comprehensive traffic intensity of all units, set the traffic intensity threshold, filter high-intensity network units, aggregate continuously connected high-intensity units according to the road network topology connectivity, and generate a candidate traffic corridor set.

[0145] Furthermore, the process of generating the candidate corridor set includes:

[0146] S401. Calculate the target quantile value of the comprehensive traffic intensity of all network units based on the quantile rule, and set the target quantile value as the traffic intensity threshold.

[0147] S402. Traverse all network units in the network unit graph and select high-intensity network units whose comprehensive traffic intensity is greater than or equal to the traffic intensity threshold.

[0148] S403. Based on the spatial adjacency relationship of highway network topology data, perform aggregation operation on high-intensity network units with direct connectivity, combine continuously connected high-intensity network units into a single candidate passage corridor, and generate a candidate passage corridor set by summarizing all continuously connected combinations.

[0149] In practical implementation, during the high-intensity network unit connectivity aggregation stage, the passage strength threshold is first determined based on the quantile rule. Extract the comprehensive traffic intensity dataset of all basic network units. ,Will The sorted sequence is obtained by arranging in ascending order. .

[0150] In this embodiment, the 80th quantile is used as the target quantile to calculate the corresponding position of the quantile. ,like If it is an integer, then ;like If it is a decimal, then In the formula, This is the floor function. It is a rounding function. This is the final traffic strength threshold.

[0151] Traverse all basic network units in the network unit graph and filter out those that meet the requirements. High-intensity network units, generating high-intensity network unit sets. Based on the spatial adjacency relationship of the road network connectivity adjacency matrix, for The high-intensity network unit performs connectivity aggregation operation, initializes the candidate corridor set to an empty set, and traverses... For each high-strength network unit that has not been aggregated, starting with that unit, recursively traverse all directly connected units that belong to the adjacency matrix. The adjacent units are processed until no new connected high-strength units are added. All continuously connected high-strength network units are combined into a candidate passageway. All units in the combination are marked as aggregated, and the candidate passageway is added to the candidate set.

[0152] After the traversal is complete, a complete set of candidate corridors is generated. ;in, The total number of candidate corridors, the first Candidate passage corridors The corresponding continuous high-intensity network unit sequence is denoted as , This represents the number of basic network units contained in the corridor.

[0153] S5. Calculate the three core evaluation indicators of continuity, traffic intensity and time period stability for each candidate corridor to construct an evaluation matrix. After dimensionless processing and objective weighting, calculate the comprehensive evaluation score, complete the level division according to the preset level threshold, and output the leveled list of new energy vehicle traffic corridors.

[0154] Furthermore, the process of generating the corridor evaluation matrix includes:

[0155] S501. For each candidate passageway in the candidate passageway set, calculate three core evaluation indicators. The first indicator is the continuity indicator, which is calculated based on the number of network units contained in the passageway and the connectivity integrity. The second indicator is the passage intensity indicator, which is calculated based on the mean and sum of the comprehensive passage intensity of all network units in the passageway. The third indicator is the time-period stability indicator, which is calculated based on the coefficient of variation of the time-period passage intensity of all network units in the passageway.

[0156] S502. Arrange the three core evaluation indicators corresponding to each candidate corridor in dimensional order to generate an indicator vector for each candidate corridor. After summing the indicator vectors of all candidate corridors, generate a corridor evaluation matrix with multiple indicator labels.

[0157] In practical implementation, during the construction of the multi-dimensional evaluation system for transit corridors, each candidate transit corridor in the candidate transit corridor set is considered. We calculate three core evaluation indicators and construct a multi-dimensional evaluation system.

[0158] The first item is the continuity indicator. The formula for characterizing the spatial connectivity integrity and continuous coverage length of a corridor is as follows:

[0159] ;

[0160] In the formula, For the first The number of high-intensity network units contained in each candidate passageway; This represents the total number of network units in the complete highway segment to which the corridor is located, indicating the connectivity ratio of the corridor within its segment. It is a natural logarithmic function, used to nonlinearly amplify the number of units, highlighting the advantages of long-distance continuous corridors; The larger the value, the better the spatial continuity of the corridor and the longer its coverage length.

[0161] The second item is the traffic intensity index. The formula for calculating the overall traffic volume of new energy vehicles in the corridor is as follows:

[0162] ;

[0163] In the formula, For the first The corridor contains the first The overall traffic strength of each basic network unit; The weighting coefficient for the mean term is given in this embodiment. , The weighting coefficients of the summation term, the mean term represents the average traffic intensity of units within the corridor, and the summation term represents the overall traffic volume of the corridor. The combination of the two avoids the distortion of indicators caused by the length deviation of a single corridor. The larger the value, the higher the popularity of new energy vehicles in the corridor.

[0164] The third item is the time-period stability index. This is used to characterize the fluctuation of traffic intensity within the corridor throughout the entire time period, and the calculation formula is:

[0165] ;

[0166] In the formula, For the first The corridor contains the first The coefficient of variation of the time-division traffic intensity of each basic network unit is calculated using the following formula: ;in, The standard deviation of the time-of-day traffic intensity for this unit is given. This is the arithmetic mean of the time-of-day traffic intensity for the entire unit. ,but ; The range of values ​​is The larger the value, the better the stability of the corridor's traffic intensity throughout the entire time period, and the smaller the fluctuation.

[0167] The three core evaluation indicators for each candidate corridor are arranged in dimensional order to generate the corresponding [number]. Indicator vector of a corridor Summarize the index vectors of all P candidate corridors to generate a P×3 dimension corridor evaluation matrix X, expressed as:

[0168]

[0169] .

[0170] Furthermore, the process for generating the list of corridors for new energy vehicles is as follows:

[0171] S601. The graded evaluation model is the entropy weight-TOPSIS comprehensive evaluation model. First, data standardization processing is performed based on the corridor evaluation matrix to eliminate the differences in the dimensions of each indicator. Then, the objective weights corresponding to each evaluation indicator are calculated based on the entropy weight method. The weighted standardization matrix of each candidate passage corridor is calculated in combination with the standardized matrix.

[0172] S602. Determine the positive and negative ideal solutions based on the weighted normalization matrix, calculate the Euclidean distance between each candidate corridor and the positive and negative ideal solutions, and obtain the comprehensive proximity of each candidate corridor.

[0173] S603. Based on the proximity classification threshold, classify each candidate passage corridor into different levels and output the classified list of new energy vehicle passage corridors.

[0174] The steps for classifying candidate passageways based on proximity grading thresholds are as follows:

[0175] S611. Set three levels of proximity thresholds, namely the first threshold, the second threshold, and the third threshold, wherein the first threshold is greater than the second threshold, and the second threshold is greater than the third threshold;

[0176] S612. Candidate passage corridors with a comprehensive proximity score greater than or equal to the first threshold are classified as core backbone corridors, those with a comprehensive proximity score between the first and second thresholds are classified as important support corridors, and those with a comprehensive proximity score between the second and third thresholds are classified as potential cultivation corridors.

[0177] S613. Generate a tiered list of new energy vehicle passage corridors, sorted by level, including the geographical scope of the corridor, comprehensive proximity, and indicators of various dimensions.

[0178] In practical implementation, during the comprehensive grading output stage based on the entropy weight-TOPSIS model, the entropy weight-TOPSIS comprehensive evaluation model is used to perform objective grading of the corridors. First, the corridor evaluation matrix X is subjected to minmax positive standardization to eliminate the dimensional differences of each indicator. The standardization formula is as follows:

[0179] ;

[0180] In the formula, , , respectively corresponding to the first The corridor, the Evaluation indicators; The first in the original evaluation matrix Line number The original index values ​​of the column; For the first The minimum value of the entire sample for each indicator. For the first The maximum value of the entire sample for each indicator; The standardized indicator value has a range of values. ,like ,but After standardization, Dimensional normalization matrix .

[0181] The objective weights of each evaluation index are calculated based on the entropy weight method. First, the objective weights of the indexes are calculated. The first item under the indicator The proportion of indicators for each corridor The calculation formula is:

[0182] ;

[0183] In the formula, if ,but .

[0184] Calculation based on index weighting Entropy value of the item index The calculation formula is:

[0185] ;

[0186] In the formula, Let be the natural logarithm function, if Then define , The range of values ​​is The smaller the entropy value, the greater the dispersion of the indicator, the more information it provides, and the higher its corresponding weight.

[0187] Based on entropy calculation, the first Coefficient of difference of the items Finally, the calculation of the first Objective weight of each indicator The calculation formula is:

[0188] ;

[0189] In the formula, satisfying , The range of values ​​is Finally, the weight vector is obtained. .

[0190] Calculate the weighted normalization matrix Z based on the normalization matrix and weight vector, and the matrix elements... ,Right now ;in, For Hadamard products, for The row weight vector repeating matrix, each row is... Dimensions and Consistent.

[0191] The positive and negative ideal solutions are determined based on the weighted normalization matrix. The vector formed by the weighted and standardized maximum values ​​of each indicator is expressed as follows: ;in, Negative ideal solution The vector formed by the weighted and standardized minimum values ​​of each indicator is expressed as follows: ;in, , .

[0192] Calculate the Euclidean distance between each candidate corridor and the positive and negative ideal solutions. The Euclidean distance between the corridor and the ideal solution The calculation formula is:

[0193] ;

[0194] No. Euclidean distance between the corridor and the negative ideal solution The calculation formula is:

[0195] ;

[0196] In the formula, The smaller the value, the closer the corridor is to the ideal solution; The larger the value, the further the corridor is from the negative ideal solution.

[0197] The overall proximity of each candidate passageway is calculated based on Euclidean distance. The calculation formula is:

[0198] ;

[0199] In the formula, The range of values ​​is , The closer The higher the overall quality of the corridor, the higher its grade.

[0200] A three-level proximity threshold is set to implement the level division. In this embodiment, the first threshold is... Second threshold The third threshold ,satisfy The threshold can be dynamically adjusted according to the needs of road network operation. The classification rule is: comprehensive proximity. The candidate passageways are divided into core backbone corridors; These are divided into important support corridors; These are divided into potential cultivation corridors; Those that are removed from the final list. Finally, the corridors are sorted in the order of core backbone corridors, important support corridors, and potential development corridors to generate a graded list of new energy vehicle traffic corridors. The list includes a unique number for each corridor, the road segment it belongs to, the geographical start and end nodes, a list of basic network units it contains, the total mileage, the comprehensive proximity, the values ​​of the three core evaluation indicators, and complete information on the corresponding grade.

[0201] The method for identifying and classifying new energy vehicle traffic corridors also includes: setting an update cycle, acquiring new highway ETC toll data within the corresponding cycle when each update cycle arrives, performing the same cleaning and new energy vehicle identification operations as S1 on the new highway ETC toll data, and generating new valid new energy vehicle traffic records.

[0202] Based on the newly added valid new energy vehicle passage records, operations S2 to S6 are executed sequentially to generate an updated list of graded new energy vehicle passage corridors. At the same time, historical graded results are compared and verified, and warning information on changes in corridor grade is output.

[0203] In practical implementation, a periodic dynamic update and grade change early warning mechanism is set up, with a fixed update cycle. This embodiment adopts a weekly update cycle. At the end of each update cycle, the newly added highway ETC passage flow data within the corresponding cycle is obtained. The newly added data is subjected to the same rule-based cleaning, time logic verification, field integrity verification, and new energy vehicle identification operations as described above, generating newly added valid new energy vehicle passage records. Based on the newly added valid new energy vehicle passage records, path reconstruction, network unit traffic intensity calculation, high-intensity unit aggregation, multi-dimensional evaluation, and comprehensive grading operations are performed sequentially to generate an updated graded new energy vehicle passage corridor list. At the same time, the updated list is compared and verified with the historical grading results of the previous cycle. If the grade of a corridor changes by more than one grade, or if a new core backbone corridor is added or an existing core backbone corridor is downgraded or removed, an early warning information on the corridor grade change is output and pushed to the road network operation and management department, providing data support for the optimization of energy replenishment facility layout, differentiated road network management, and adjustment of emergency support plans.

[0204] This embodiment is based on measured data of the expressway network in southern Jiangsu Province. The method of the present invention is compared with three existing mainstream traditional methods to quantitatively verify the improvement of the technical effect of the present invention. The test dataset is consistent with the data source of the specific embodiment below.

[0205] 1. Compare with the test plan

[0206] Three mainstream methods from existing technologies were selected as a control group:

[0207] Control group A (traditional path restoration method): The conventional path restoration method adopts the shortest path constraint + single-layer average speed verification, without energy replenishment time redundancy adaptation design;

[0208] Control group B (traditional feature quantification method): Only the absolute traffic volume of new energy vehicles is used as a single indicator to represent traffic popularity, without traffic ratio weighting design;

[0209] Control group C (traditional corridor identification method): uses fixed thresholds set by human experience to screen road segments, without standardized continuous unit recursive aggregation rules.

[0210] 2. Comparison of Quantitative Technique Effects Data

[0211] Comparison Dimensions Method of the present invention Comparison method Improved technical performance Path restoration accuracy 94.6% Control group A: 72.3% The accuracy rate improved by 22.3 percentage points, and the effective path recognition rate in the power replenishment scenario increased from 31.2% to 92.8%. New Energy Corridor Identification Deviation 11.2% Control group B: 38.7% The overlap between the actual high-frequency traffic sections and the actual routes used by new energy vehicles decreased by 27.5 percentage points, and the recall rate for identifying high-frequency routes specifically for new energy vehicles improved by 41.6%. Average number of consecutive units in a corridor 8.7 Control group C: 3.2 The continuity of corridor space improved by 171.9%, and the proportion of fragmented road sections decreased from 68.4% to 12.7%. Corridor time stability Coefficient of variation: 0.11 Control group C: Coefficient of variation 0.19 Traffic intensity fluctuations decreased by 42.1% during different time periods, resulting in a significant improvement in traffic stability throughout the entire time period. Hit rate of energy replenishment facility layout 89.2% Control group C: 56.8% The utilization rate of energy replenishment facilities within the core corridor increased by 32.4 percentage points, and the proportion of ineffective facilities decreased by 58.3%.

[0212] 3. Core Effects Description

[0213] This invention solves the core problem of traditional methods misjudging long-duration, long-mileage effective paths caused by recharging as abnormal data by using recharging time redundancy and two-level spatiotemporal verification, thus significantly improving the accuracy of path reconstruction for new energy vehicles.

[0214] This invention accurately identifies high-frequency traffic sections for new energy vehicles by characterizing the traffic intensity of new energy vehicles, avoiding the identification bias caused by the traditional method's "emphasis on traffic volume and neglect of proportion". At the same time, it realizes the automatic generation of continuous corridors through a recursive aggregation mechanism, solving the problem of corridor fragmentation in the traditional method.

[0215] The hierarchical corridors generated by this invention significantly improve the matching degree with the actual energy replenishment needs of new energy vehicles, and can directly support the precise layout of high-speed new energy replenishment facilities, greatly reducing the construction cost of ineffective facilities.

[0216] Example Implementation Basic Data Description

[0217] 1. Data Sources and Scale

[0218] The test scope of this embodiment is the core expressway network in the southern Jiangsu region of Jiangsu Province. The specific parameters are as follows:

[0219] The road network covers the core sections of the G2 Beijing-Shanghai Expressway, G42 Shanghai-Chengdu Expressway, and G1522 Changtai Expressway, including 127 toll stations, 89 interchange hubs, and 176 basic network units between adjacent interchanges (the smallest analysis unit of this method), with a total road network mileage of 1286 kilometers.

[0220] ETC transaction data: Selected full ETC network toll transaction data from October 1, 2025 to October 7, 2025 (peak period of new energy vehicle travel on highways during the National Day holiday), a total of 7 calendar days, with a cumulative total of 12,864,237 original passage records;

[0221] Supporting basic data: 2025 version of expressway network vector topology data released by Jiangsu Provincial Department of Transportation, and Jiangsu Province new energy vehicle type code comparison table based on the Ministry of Transport's "Vehicle Classification for Toll Roads" (JT / T489-2019).

[0222] 2. Data cleaning results

[0223] For 12,864,237 original ETC transaction records, the double verification and new energy vehicle identification were performed in step S1. The results are as follows:

[0224] Time logic verification: 327,842 abnormal records were removed, accounting for 2.55% of the total original data, where the exit passage time was earlier than the entrance passage time or the passage duration exceeded the reasonable passage range of OD.

[0225] Field integrity verification: 186,429 invalid records, accounting for 1.45% of the total original data, were removed because the core fields (vehicle type code, entrance / exit toll station, and passage time) had empty values ​​or invalid codes.

[0226] Valid passage records for all vehicle types after cleaning: 12,349,966;

[0227] New energy vehicle identification results: Based on the new energy vehicle type coding reference table, 2,148,743 valid new energy vehicle passage records were matched and extracted, accounting for 17.41% of the total number of records of all vehicle types after cleaning, which will serve as the basis dataset for subsequent analysis.

[0228] Example 1: Complete Example of Path Reconstruction

[0229] This example demonstrates the complete path reconstruction process of the present invention for a single new energy vehicle passage record, while also comparing it with the processing deficiencies of traditional methods.

[0230] 1. Basic parameters

[0231] For new energy passenger vehicles, the entrance toll station is Suzhou New District Toll Station, and the corresponding interchange is Suzhou New District Interchange (starting point node). The exit toll station is Wuxi East Toll Station, and the corresponding access interchange is Wuxi East Interchange (endpoint node). );

[0232] Entry time: 2025-10-02 08:12:00 =8.2h), Exit time 2025-10-02 11:42:00 ( =11.7h), actual travel time ΔT=3.5h;

[0233] The theoretical shortest path mileage of this OD pair is calculated using Dijkstra's algorithm. =92km (G2 Beijing-Shanghai Expressway main line).

[0234] 2. Results of traditional methods

[0235] The traditional shortest path + single-layer speed verification method calculates the reasonable travel time range for this OD pair as [92 / 120≈0.77h, 92 / 60≈1.53h]. The actual travel time of 3.5h far exceeds the upper limit of the range, so it is directly judged as an abnormal record and removed, and the path restoration fails.

[0236] 3. Path reconstruction process of the present invention

[0237] Step 1: Define the mileage constraint interval for candidate paths

[0238] OD to theoretical shortest path mileage =92km;

[0239] Maximum legal driving speed 120km / h × actual travel time 3.5h = 420km;

[0240] The maximum time redundancy coefficient is set to 1.5, covering the incremental time of new energy vehicles stopping for refueling and slowing down.

[0241] Step 2: Generate a set of acyclic candidate paths

[0242] The Yen's K Shortest Paths algorithm was used, with K=10, to generate 10 acyclic simple paths from the start point to the end point. After filtering by mileage constraint intervals, 3 candidate paths that meet the constraint [92km, 420km] were retained, as follows:

[0243] Candidate path number Path Total distance (km) Path 1 Suzhou New District Interchange → G2 Beijing-Shanghai Expressway Mainline → Wuxi East Interchange 92 Path 2 Suzhou New District Interchange → G1522 Changtai Expressway → G42 Shanghai-Chengdu Expressway → Wuxi East Interchange 118 Path 3 Suzhou New District Interchange → S9 Su-Shao Expressway → G42 Hu-Cheng Expressway → Wuxi East Interchange 136

[0244] Step 3: Two-level spatiotemporal verification to filter the optimal path.

[0245] Level 1 Verification (Speed ​​Range Verification): Calculate the theoretical average driving speed of each candidate path and verify whether it falls within the legal speed range [60km / h, 120km / h]. The reasonable refueling stop time is set to max(2h, 15 minutes per 100km) = max(2h, 92 / 100 × 15min) = 2h, and the pure driving time is 3.5h - 2h = 1.5h.

[0246] The average speed along route 1 is approximately 61.3 km / h (92 km / 1.5 h), which meets the requirements for the interval.

[0247] The average speed along route 2 is approximately 78.7 km / h (118 km / 1.5 h), which meets the requirements for the section.

[0248] The average speed along route 3 is approximately 90.7 km / h (136 km / 1.5 h), which meets the requirements for the section.

[0249] All three paths passed the first-level verification.

[0250] Secondary verification (duration coverage verification): Verify whether the theoretical shortest travel time of the candidate route and the longest reasonable travel time of the superimposed energy replenishment stop cover the actual travel time of 3.5 hours.

[0251] Path 1: Theoretical shortest travel time = 92 / 120 ≈ 0.77h, longest reasonable travel time = 92 / 60 + 2h ≈ 1.53h + 2h = 3.53h, 0.77h ≤ 3.5h ≤ 3.53h, complete coverage, passed verification;

[0252] Route 2: Theoretically shortest travel time = 118 / 120 ≈ 0.98h, longest reasonable travel time = 118 / 60 + 2h ≈ 1.97h + 2h = 3.97h, 0.98h ≤ 3.5h ≤ 3.97h, passed verification;

[0253] Route 3: Theoretically shortest travel time = 136 / 120 ≈ 1.13h, longest reasonable travel time = 136 / 60 + 2h ≈ 2.27h + 2h = 4.27h, 1.13h ≤ 3.5h ≤ 4.27h, passed verification.

[0254] Step 4: Determining the optimal path

[0255] Calculate the spatiotemporal matching degree of the three paths, using the historical average driving speed of the default region. =90km / h, the matching degree formula is: ;

[0256] Path 1 matching degree: ;

[0257] Path 2 matching degree: ;

[0258] Path 3 matching degree: ;

[0259] Based on the common behavioral characteristics of high-speed travel, when the matching degree is close, the shortest route is selected first, and route 1 is finally determined as the optimal travel route.

[0260] Step 5: Path Decomposition and Sequence Generation

[0261] The optimal path is decomposed into an ordered sequence of basic network units: Suzhou New District Interchange - Dongqiao Interchange Unit, Dongqiao Interchange - Shuofang Hub Unit, and Shuofang Hub - Wuxi East Interchange Unit; a total of 3 continuous basic network units. The path reconstruction of a single record is completed and included in the full set of unit passage sequences.

[0262] Example 2: Complete example of corridor generation

[0263] This example, based on the full set of cell access sequences, fully demonstrates the process of high-intensity cell screening and continuous corridor aggregation.

[0264] 1. Basic parameters

[0265] Analysis objects: 8 continuous basic network units of the Suzhou-Wuxi section of the G2 Beijing-Shanghai Expressway, numbered E1-E8;

[0266] Time window setting: A single day of 24 hours is divided into 24 equal-length windows of 1 hour each. The time-sharing traffic intensity of each unit is calculated, and the average value of the entire time period is taken to obtain the comprehensive traffic intensity.

[0267] Traffic intensity calculation formula: , where α=0.6 and β=0.4.

[0268] 2. Overall Traffic Intensity Results for Each Unit

[0269] Network Unit Number Corresponding road section Comprehensive traffic intensity value E1 Suzhou New District Interchange - Dongqiao Interchange 0.82 E2 Dongqiao Interchange - Tong'an Interchange 0.78 E3 Tong'an Interchange - Shuofang Hub 0.85 E4 Shuofang Hub - Sunan Shuofang Airport Interchange 0.76 E5 Sunan Shuofang Airport Interchange - Wuxi New District Interchange 0.88 E6 Wuxi New District Interchange - Wuxi East Interchange 0.81 E7 Wuxi East Interchange - Wuxi North Interchange 0.72 E8 Wuxi North Interchange - Jiangyin South Interchange 0.69

[0270] 3. High-intensity unit screening

[0271] The 80th percentile of the comprehensive traffic intensity of all road network units is taken as the traffic intensity threshold, and the threshold is calculated. .

[0272] Screening rules: Units with a comprehensive traffic intensity ≥ 0.75 are considered high-intensity units. The final high-intensity units selected are: E1, E2, E3, E4, E5, and E6. E7 and E8 are below the threshold and are excluded.

[0273] 4. Recursive aggregation of continuous units

[0274] Based on the directed adjacency matrix of the road network, perform a recursive aggregation operation:

[0275] Starting with the unaggregated E1 as the initial node, traverse the adjacency matrix. The endpoints of E1 and the starting points of E2 are directly connected, and E2 is a high-intensity unit, so it is included in the aggregation sequence.

[0276] With E2 as the current node, the endpoint of E2 is directly connected to the starting point of E3, and E3 is a high-intensity unit, so it is included in the aggregation sequence;

[0277] By recursively traversing the sequence, all of E1-E6 eventually become continuous, high-strength units with no new units that can be aggregated.

[0278] One candidate traffic corridor, numbered C1, is generated. The corridor unit sequence is [E1,E2,E3,E4,E5,E6]. The corresponding road segment is the Suzhou New District Interchange to Wuxi East Interchange section of the G2 Beijing-Shanghai Expressway, with a total length of 92km and containing 6 continuous basic network units.

[0279] Example 3: Complete example of corridor classification results

[0280] This example fully demonstrates the multi-dimensional evaluation and grading process for the candidate corridors generated by aggregation.

[0281] 1. Basic parameters for classification

[0282] Evaluation indicators: three core indicators: continuity, traffic intensity, and time-period stability;

[0283] Evaluation model: The entropy weight-TOPSIS comprehensive evaluation model (preferred implementation method) is adopted.

[0284] The threshold values ​​are: first threshold T1 = 0.8, second threshold T2 = 0.5, and third threshold T3 = 0.2.

[0285] Grading rules: A comprehensive proximity score of ≥0.8 is a core backbone corridor, 0.5≤proximity score<0.8 is an important support corridor, 0.2≤proximity score<0.5 is a potential development corridor, and <0.2 is removed from the list.

[0286] 2. Calculation of core indicators (taking candidate corridor C1 as an example)

[0287] Continuity Index X1: The corridor comprises 6 network units, and the total number of units corresponding to the G2 Beijing-Shanghai Expressway is 8. The calculation formula is as follows:

[0288] ;

[0289] Traffic intensity index X2: The average comprehensive traffic intensity of units within the corridor is approximately 0.817, the total intensity is 4.9, and γ is 0.5. The calculation formula is as follows: ;

[0290] Time-of-day stability index X3: The mean coefficient of variation of time-of-day traffic intensity for each unit within the corridor is 0.12, calculated using the following formula: ;

[0291] 3. Comprehensive Evaluation and Grading Results

[0292] After min-max positive standardization, entropy weight calculation of objective weights for indicators (continuity 0.32, traffic intensity 0.45, time-period stability 0.23), weighted standardization matrix calculation, and Euclidean distance calculation for positive and negative ideal solutions, the final comprehensive proximity score of candidate corridor C1 is 0.86. Since 0.86 ≥ 0.8 (the first threshold), corridor C1 is classified as a core backbone corridor.

[0293] 4. Other examples of grading

[0294] Candidate corridor number Corresponding road section Overall relevance Classification C2 G42 Shanghai-Chengdu Expressway, Wuxi East Interchange - Changzhou North Interchange Section 0.62 Important support corridor C3 G1522 Changtai Expressway Suzhou City Interchange - Wujiang Interchange Section 0.35 Potential Cultivation Corridor C4 S9 Su-Shao Expressway Tong'an Interchange - Dongqiao Interchange Section 0.17 Removed from the final list

[0295] Finally, the corridors are sorted by level to generate a tiered list of new energy vehicle passage corridors, which includes corridor number, road segment, start and end nodes, total mileage, comprehensive proximity, core indicators, and corresponding level.

[0296] Example 4: Periodic Dynamic Update and Early Warning Mechanism

[0297] This embodiment sets a fixed weekly update cycle. When the cycle reaches October 8, 2025, the newly added ETC transaction data from October 1 to 7, 2025 is obtained, and steps S1-S5 are repeated to generate an updated classification list. The updated result is compared and verified with the historical classification results of the fourth week of September 2025, and it is found that:

[0298] This time, one new core backbone corridor has been added (G4221 Shanghai-Wuhan Expressway Changzhou-Zhangjiagang section).

[0299] The original core backbone corridor, the Taicang section of the G15 Shenhai Expressway, has had its overall proximity score reduced to 0.71, and has been downgraded to an important support corridor, a change of more than one level.

[0300] In response to the above changes, the system automatically generates early warning information on changes in corridor levels and pushes it to the road network operation and management department to provide data support for optimizing the layout of energy replenishment facilities and differentiated management of the road network.

[0301] Based on the preferred embodiments of the present invention described above, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for identifying and classifying a new energy vehicle passageway based on ETC pass data, characterized in that, Includes the following steps: S1. Obtain the original ETC transaction data of the highway containing vehicle type code, entrance and exit toll station and corresponding passage time fields. After double verification of time logic and field integrity, abnormal and invalid records are removed. Valid new energy vehicle passage records are extracted based on the new energy vehicle type code lookup table. S2. Obtain highway network topology data containing interconnected nodes and basic network unit attributes between adjacent interconnections, and construct a directed adjacency matrix of the road network; take the OD corresponding to the interconnection of a single passage record as the start and end point, and combine the refueling time redundancy of new energy vehicles to delineate the candidate path mileage constraint interval, generate acyclic candidate paths that meet the interval requirements, screen the optimal driving path through two levels of spatiotemporal speed and duration verification, decompose it into an ordered basic network unit sequence, and summarize to generate a full set of unit passage sequences; S3. Statistically calculate the traffic volume and traffic ratio of new energy vehicles in each network unit according to the same time window, calculate the time-sharing traffic intensity by weighting, take the average value of the whole time period to generate the comprehensive traffic intensity of each unit, and construct a network unit map with intensity labeling. S4. Based on the quantile of the comprehensive traffic intensity of all units, set the traffic intensity threshold, filter high-intensity network units, aggregate continuously connected high-intensity units according to the road network topology connectivity, and generate a candidate traffic corridor set. S5. Calculate the three core evaluation indicators of continuity, traffic intensity and time period stability for each candidate corridor to construct an evaluation matrix. After dimensionless processing and objective weighting, calculate the comprehensive evaluation score, complete the level division according to the preset level threshold, and output the leveled list of new energy vehicle traffic corridors.

2. The ETC toll data-based new energy vehicle toll corridor identification and classification method according to claim 1, characterized in that: In S1, the time logic check is to remove records where the exit passage time is earlier than the entrance passage time and the passage duration exceeds the reasonable passage range of OD; the reasonable passage range of OD is calculated based on the shortest path mileage of OD and the statutory speed limit of the highway; the new energy vehicle type code lookup table is constructed according to the toll road vehicle classification standard of the Ministry of Transport and the unified coding rules of new energy vehicles of each province's network toll collection system.

3. The method for identifying and classifying new energy vehicle traffic corridors based on ETC traffic data according to claim 1, characterized in that: In S2, the lower limit of the mileage constraint interval is the mileage of the theoretical shortest path from OD to 0, and the upper limit is the product of the highest legal driving speed and the actual travel time. The refueling time redundancy is reflected by the maximum time redundancy coefficient, which ranges from 1.2 to 1.8 and is used to cover the time increment of refueling stops and slow driving of new energy vehicles.

4. The method for identifying and classifying new energy vehicle traffic corridors based on ETC traffic data according to claim 1, characterized in that: In S2, the first-level check of the two-level spatiotemporal verification is that the theoretical average driving speed of the candidate path falls within the legal speed range of the OD pair. The second-level check is that the longest reasonable driving time of the candidate path's theoretical shortest driving time and the superimposed reasonable dwell time for energy replenishment covers the actual travel time. The reasonable dwell time for energy replenishment is the maximum value of a fixed 2 hours and a dwell time of 15 minutes per 100 kilometers.

5. The method for identifying and classifying new energy vehicle traffic corridors based on ETC traffic data according to claim 1, characterized in that: In S3, the formula for calculating time-sharing traffic intensity is: ； In the formula, For the first The network unit in the first Time-based traffic intensity for each time window This represents the traffic volume of new energy vehicles after the min-max standardization. The proportion of new energy vehicles on the road. This is a weighting coefficient for the traffic volume of new energy vehicles. This is the weighting coefficient for the proportion of new energy vehicles, and , The value ranges from 0.5 to 0.

7. The value ranges from 0.3 to 0.

5.

6. The method for identifying and classifying new energy vehicle traffic corridors based on ETC traffic data according to claim 1, characterized in that: In S4, the quantile is taken as the 80th quantile value; the aggregation operation is a recursive aggregation, the rule is: traverse the high-intensity units that have not been aggregated, take the unit as the initial node, recursively traverse all directly connected high-intensity units based on the directed adjacency matrix until there are no new units, and combine the continuously connected units into a single candidate passageway.

7. The ETC toll data-based new energy vehicle toll corridor identification and classification method according to claim 1, characterized in that: In S5: The continuity index is calculated based on the number of network units contained in the corridor and the connectivity integrity of the road segment to which it belongs; The traffic intensity index is calculated based on the average and sum of the comprehensive traffic intensity of all network units within the corridor; The time-period stability index is calculated based on the coefficient of variation of the time-based traffic intensity of all network units within the corridor.

8. The ETC toll data-based new energy vehicle toll corridor identification and classification method according to claim 1, characterized in that: In S5, the steps for objective weight assignment and comprehensive evaluation using the entropy-weight-TOPSIS comprehensive evaluation model are as follows: S501. The evaluation matrix is ​​subjected to dimensionless processing based on the min-max positive standardization method to obtain the standardized matrix; S502. Calculate the objective weights of each evaluation index based on the entropy weight method, and generate a weighted standardized matrix by combining the standardized matrix. S503. Determine the positive and negative ideal solutions for each indicator, calculate the Euclidean distance between each candidate corridor and the positive and negative ideal solutions, and calculate the comprehensive proximity score based on the Euclidean distance. S504. Based on the preset three-level proximity threshold, complete the level division and output the level list.

9. The method for identifying and classifying new energy vehicle traffic corridors based on ETC traffic data according to claim 8, characterized in that: The three-level proximity thresholds are the first threshold, the second threshold, and the third threshold, which decrease sequentially. Those with a comprehensive proximity greater than or equal to the first threshold are classified as core backbone corridors, those with a first threshold greater than or equal to the comprehensive proximity and the second threshold are classified as important support corridors, those with a second threshold greater than or equal to the comprehensive proximity and the third threshold are classified as potential cultivation corridors, and those with a proximity lower than the third threshold are removed from the list.

10. The ETC toll data-based new energy vehicle toll corridor identification and classification method according to claim 1, characterized in that: The method also includes a periodic dynamic update mechanism: a fixed update cycle is set, and when the cycle arrives, new ETC transaction data is obtained, and steps S1 to S5 are repeated to generate an updated classification list; the update results are compared and verified with the historical results of the previous cycle, and if the corridor level changes by more than one level, a new core backbone corridor is added, or an existing core backbone corridor is removed, a level change warning message is output.

Images