Urban road network key path automatic identification method based on trajectory segmentation

Abstract

The invention discloses an urban road network key path automatic identification method based on trajectory segmentation, and solves the problems that a node-based road network key path identification method is insufficient in consideration of dynamic traffic flow and a key road section-based road network key path identification method is low in credibility. The method comprises the following steps: 1) acquiring a complete continuous driving track of a vehicle according to checkpoint data; 2) a vehicle path decision chain is formed, and coding and traffic flow statistics are carried out on the vehicle path decision chain; and 3) based on the connection rule of the vehicle path decision chain and a criticality calculation formula, forming a critical path in a road network, comprising the following steps: (1) giving the connection rule between the vehicle path decision chains; (2) proposing a criticality calculation formula of a vehicle path decision chain; (3) determining the value range of the criticality; (4) screening a starting intersection and calculating the criticality of all vehicle path decision chains in the road network; (5) screening and connecting vehicle path decision chains forming connection paths; and (6) forming a critical path in the road network.

Description

Technical Field

[0001] The present invention relates to a method in the technical field of urban intelligent traffic management, and in particular to a method for automatically identifying key paths in an urban road network based on trajectory segmentation. Background Art

[0002] In signal coordination and control, critical paths are defined as a series of major routes that play a decisive role in the overall operational efficiency of a transportation network. These routes are closely associated with heavy traffic flows, and traffic along these critical paths often receives a greater allocation of green light time at intersections. Therefore, improving the overall efficiency of a road network requires signal coordination and optimization of these critical paths, aiming to achieve optimal control results for the network or arterial route. Therefore, establishing an efficient and accurate method for identifying critical paths is essential to ensure the smooth implementation of signal coordination and control.

[0003] Existing research on critical path identification primarily focuses on key nodes and key links. Critical node analysis is typically based on complex network theory, focusing on issues in graph theory and graph data mining. Node importance is defined as a significant characteristic reflecting the degree of connectivity differentiation of a node in a network. Attributes such as degree, betweenness, and eigenvectors are used as core metrics for evaluating critical nodes. Simulating the impact of node deletion on overall network connectivity evaluates the node's importance and impact within the complex network. Research on key nodes primarily focuses on static network distances, neglecting dynamic traffic flow conditions. Research on key links focuses on three distinct aspects: defining key links based on vulnerability, reliability, or network topology; optimizing network reliability based on key links; and identifying key links. This approach comprehensively considers link queue capacity and cross-sectional capacity, employs metrics such as link connectivity, betweenness, and saturation, and calculates failure probabilities based on reliability theory to identify key links that contribute to significant network vulnerability. This approach verifies the potential for destructive effects on the overall operational efficiency of the network if a key link is destroyed. In the research conducted around key sections, the corresponding theoretical models and evaluation indicators vary greatly due to different research perspectives, resulting in significant discrepancies between the research results and the actual situation of the road network, that is, the credibility of the critical paths formed is not high.

[0004] To address the shortcomings of existing research methods, the present invention creatively proposes a method for automatic identification of critical paths based on trajectory segmentation. This method quantifies the importance of vehicle path decision chains (VPDCs) using a criticality calculation formula, laying the foundation for the formation of critical paths. The method, described in the present invention, filters VPDCs based on their criticality before connecting them to form a critical path. Compared to methods that first connect all local paths to form candidate critical paths and then filter them to obtain the critical path, this method significantly reduces workload and improves the overall efficiency of critical path identification. Furthermore, the proposed method uses criticality as the basis for filtering VPDCs, cleverly avoiding a major drawback of existing research methods, namely, relying on field traffic flow data to calculate the average turn ratio, perform traffic flow distribution at downstream intersections, and identify critical paths based on traffic flow distribution at downstream intersections. This traditional method easily leads to inconsistent path flow estimates with actual conditions, thereby affecting the accuracy of critical path estimation and ultimately adversely affecting signal timing design. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the problems in the prior art of insufficient consideration of dynamic traffic flow conditions by the node-based road network critical path identification method and low credibility of the road network critical path identification method based on key sections, and provide an automatic identification method for urban road network critical paths based on trajectory segmentation.

[0006] To solve the above technical problems, the present invention is implemented by adopting the following technical solution: the method for automatically identifying key paths in urban road networks based on trajectory segmentation comprises the following steps:

[0007] 1) Obtain the complete and continuous driving trajectory of the vehicle based on the checkpoint data;

[0008] 2) Form a vehicle path decision chain and encode it and calculate traffic flow statistics;

[0009] 3) Based on the connection rules and criticality calculation formula of the vehicle path decision chain, the critical path in the road network is formed.

[0010] The steps described in the technical solution for obtaining the complete continuous driving trajectory of the vehicle based on the checkpoint data are as follows:

[0011] 1) Collect real-time vehicle passing data detected by the checkpoint video detection equipment:

[0012] x i =(num,p i,t i )

[0013] Where: i is the bayonet number, x i is the vehicle passing data at checkpoint i, num is the vehicle license plate, p is the coordinates of the vehicle's trajectory point, which is determined according to the position of the checkpoint equipment; t is the time when the vehicle arrives at the checkpoint;

[0014] 2) Through intelligent processing of vehicle passing data detected by all checkpoints in the area, the passing records of the same vehicle at different checkpoints are classified and integrated according to the vehicle license plate information to form a complete set of vehicle passing data for each vehicle:

[0015] If the num in the passing data is the same, it is the passing data of the same vehicle and is integrated into the passing point set X of the vehicle. num In, that is

[0016] X num ={x1,…,x i ,…,x n}

[0017] 3) To avoid interference of vehicle stops on path identification, the continuity of the checkpoint data is judged and screened:

[0018] This step is mainly to ensure that the complete vehicle passing data set is formed by a single trip within a certain time range, so the continuity of the checkpoint data is judged and screened;

[0019]

[0020] Where: t is the time measurement value for judging the continuity of the trajectory; It is the average distance from the stop line in a certain direction of the upstream intersection to the stop line in a certain direction of the downstream intersection of all adjacent intersections in the road network; The average speed of all valid vehicles in the road network, which can be determined based on the historical speed data of the road network area; is the average red light time in all directions of all intersections in the road network; η is the disturbance factor, which is determined according to the specific conditions of the vehicle driving process;

[0021]

[0022] Where: d y,a,b The distance from the stop line in the direction of a on the entrance road of intersection y to the stop line in the direction of b on the entrance road of intersection y+1;

[0023] Δt i =t i+1 -t i

[0024] If Δti >t, then n=i

[0025] Use the loop structure to verify Δt in sequence i , if Δt i ≤t, then i=i+1, until Δt i >t, the loop ends, and n=i.

[0026] Where: Δt i For a car from trajectory point p i To trajectory point p i+1 The time taken, n is the set of vehicles with license plate number num passing by X num The number of vehicles that have passed the test;

[0027] 4) Use ArcGIS software to perform road network map matching to obtain the complete continuous driving trajectory of the vehicle:

[0028] Using ArcGIS software, the passing points of each vehicle are set as X num ={x1,…,x i ,...,x n} is matched with the map of the road network under study to form a complete and continuous driving trajectory of each vehicle in real time.

[0029] The steps for forming a vehicle routing decision chain, encoding it, and calculating traffic flow statistics as described in the technical solution are as follows:

[0030] 1) Segmentation principle:

[0031] Taking the upstream intersection inflow and downstream intersection outflow between two intersections as a minimum segmentation unit, the complete vehicle driving trajectory is combined according to the upstream arrival and downstream outflow traffic flows of two adjacent intersections to form a vehicle path decision chain;

[0032] 2) Classification principles:

[0033] The combination of the upstream arrival traffic flow and the downstream departure traffic flow of vehicles at the intersection is used as the basis for classification, that is,

[0034] l=(i,j)

[0035] vpdc=(l,m,n)

[0036] Where: l represents the road section; i, j represent the upstream and downstream intersection numbers of road section l respectively; m represents the upstream arrival traffic flow attribute of the intersection, m∈{ZX,ZZ,YZ}; n represents the downstream departure traffic flow attribute of the intersection, n∈{ZX,ZZ,YZ};

[0037] According to the combination of m and n, the vehicle routing decision chain can be divided into 9 categories, as shown in the following table:

[0038] Table 1 Vehicle routing decision chain categories

[0039]

[0040] Explanation of symbols in Table 1: vpdc t represents the vehicle routing decision chain of category t, where t∈{1,2,…,9}; ZX represents the through traffic flow; ZZ represents the left-turn traffic flow; and YZ represents the right-turn traffic flow.

[0041] Figure 5a 、 Figure 5b and Figure 5c The vehicle routing decision chain categories shown in the figure correspond to the vehicle routing decision chain categories in Table 1. The types of vehicle routing decision chains are explained. When the inflow traffic flow attribute of the upstream intersection is straight, it can be combined with the straight, left-turn and right-turn traffic flows of the downstream intersection to form three types of vehicle routing decision chains, namely vpdc1: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is straight; vpdc2: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is left-turn; vpdc3: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is right-turn;

[0042] 3) Coding principles:

[0043] The vehicle routing decision chain code should include vehicle trajectory information, intersection number information and vehicle routing decision chain category information, that is,

[0044] v k =(k,l,vpdc t )

[0045] Where: v k represents the kth vehicle path decision chain from top to bottom into which the complete vehicle driving trajectory is divided; l indicates that the kth vehicle path decision chain belongs to road section l; vpdc t represents that the kth vehicle routing decision chain belongs to category t in the vehicle routing decision chain classification, where t∈{1,2,…,9};

[0046] 4) Statistics of each v i Flow value:

[0047] If the code v i If l and t are equal at the same time, they are considered to be the same and are recorded as The flow rate value is recorded as q l,t .

[0048] The steps for forming a critical path in a road network based on the connection rules and criticality calculation formula of the vehicle routing decision chain described in the technical solution are as follows:

[0049] 1) Provide the connection rules between vehicle routing decision chains;

[0050] 2) Propose a formula for calculating the criticality of the vehicle routing decision chain;

[0051] 3) Determine the value range of criticality;

[0052] 4) Screen the starting intersection and calculate the criticality of all vehicle routing decision chains in the road network;

[0053] 5) Screening and connecting vehicle routing decision chains that constitute the connection path;

[0054] 6) Form the critical path in the road network.

[0055] The connection rules between vehicle routing decision chains described in the technical solution refer to:

[0056] From the definition and classification principle of vehicle routing decision chains, we can see that vehicle routing decision chains cannot be connected arbitrarily, but need to follow certain rules. Figure 6a The departure actions of the vehicles in vpdc1, vpdc4, and vpdc7 are all straight-ahead, so the arrival actions of the vehicle routing decision chains connected to them downstream can only be straight-ahead. This means that the vehicle routing decision chain with the combination type of vpdc1, vpdc4, and vpdc7 can only be connected to the vehicle routing decision chain with the combination type of vpdc1, vpdc2, and vpdc3.

[0057] Based on the same principle, the connection rules of the vehicle routing decision chain are as follows:

[0058] (1) The VDRs of combination type vpdc1, vpdc4, and vpdc7 can only be connected to the VDRs of combination type vpdc1, vpdc2, and vpdc3;

[0059] (2) The vehicle routing decision chains with the combination type of vpdc2, vpdc5, and vpdc8 can only be connected to the vehicle routing decision chains with the combination type of vpdc4, vpdc5, and vpdc6;

[0060] (3) The vehicle routing decision chains of combination type vpdc3, vpdc6, and vpdc9 can only be connected to the vehicle routing decision chains of combination type vpdc7, vpdc8, and vpdc9.

[0061] The criticality calculation formula of the vehicle routing decision chain proposed in the technical solution is:

[0062]

[0063] Where: represents the criticality of the vehicle routing decision chain of category t on road segment l 20; l,t represents the flow of the vehicle routing decision chain of category t on road segment 20; q l,t,m represents the total flow of upstream inflow traffic flow m of the vehicle routing decision chain of category t on road segment l 20; l,t,n represents the total flow of downstream outgoing traffic flow n of the vehicle routing decision chain of category t on road segment l 20; represents the total traffic flow in the direction of the d1 entrance road at intersection i on road segment l 20, where d1∈{N,S,W,E}; Represents the total traffic flow in the direction of the d2 entrance road at intersection j on section l 20, where d2∈{N,S,W,E}.

[0064] The range of values for determining criticality described in the technical solution refers to:

[0065] The first term in the VDRC criticality calculation formula represents the percentage of the flow of VDRC type t on road segment l 20 to the total flow of the three types of VDRCs formed by the same upstream inflow traffic flow.

[0066] The second term represents the percentage of the upstream incoming traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total flow of the entrance road;

[0067] The third term represents the percentage of the downstream outflow traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total traffic flow of the entrance road;

[0068] It is not difficult to find that when the three percentages in the formula are all 1, the criticality value reaches the maximum value of 3, which means that there is only one type of vehicle path decision chain on the road section 20. The purpose of the automatic identification method of urban road network critical paths based on trajectory segmentation described in the present invention is to identify critical paths in urban road networks. According to expert experience and experimental verification results, the lower limit of the criticality value range is determined to be 1.2; in summary, the criticality value range is [1.2, 3].

[0069] The criticality of screening the starting intersection and calculating all vehicle path decision chains in the road network described in the technical solution refers to:

[0070] Determine the road network to be studied, use the total flow rate of the intersection as the basis for judging the importance of the intersection, sort the intersections in the road network from large to small according to their importance, and select the intersection with the largest flow rate as the starting intersection. After an intersection has been used as the starting intersection, the intersections within the six adjacent intersections on the same trunk line cannot be used as the starting intersection. It is necessary to select the intersection with the largest flow rate from these intersections as the starting intersection; then, calculate the criticality of all vehicle path decision chains in the road network using the criticality calculation formula of the vehicle path decision chain.

[0071] The screening and connection of the vehicle routing decision chains constituting the connection paths described in the technical solution refer to:

[0072] After determining the starting intersection, find the two VDRs with the highest criticality at the starting intersection. According to the connection rules between VDRs, find all VDRs that can be connected to the two VDRs with the highest criticality. Determine the criticality of all VDRs that meet the requirements. If the criticality of a VDR is within the interval [1.2,3], connect the VDR to the upstream VDR. Then, find all VDRs that can be connected to the VDR at the downstream intersection of the VDR that has just been connected. Determine the criticality of all VDRs that meet the requirements. If the criticality of a VDR is within the interval [1.2,3], connect the VDR to the upstream VDR. Repeat this process until one of the following conditions is met:

[0073] (1) The criticality of all connected vehicle routing decision chains is not within the interval [1.2,3];

[0074] (2) When the number of vehicle path decision chains in the connected path is 6;

[0075] (3) There are already two vehicle path decision chains that meet the requirements and are connected at the intersection;

[0076] Repeat the above process until there is no intersection in the road network that can be used as the starting intersection.

[0077] The key paths in the road network described in the technical solution are:

[0078] Output all connection paths, and accumulate the criticality of the vehicle path decision chain in each connection path, take the connection paths that rank in the top 60% of the cumulative sum as the critical paths in the road network and put them into the critical path set CP = {p n}middle.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] 1. The method for automatically identifying critical paths in urban road networks based on trajectory segmentation described in the present invention creatively proposes a formula for calculating the criticality of a vehicle routing decision chain (VDRC) by considering three percentages: the percentage of the flow of a VDRC of type t on a road section 20 to the total flow of three types of VDRCs formed by the same upstream incoming traffic flow; the percentage of the upstream incoming traffic flow of a VDRC of type t on a road section 20 to the total flow of the entrance road; and the percentage of the downstream outgoing traffic flow of a VDRC of type t on a road section 20 to the total flow of the entrance road. This method quantifies the criticality of VDRCs and provides data support for the subsequent formation of critical paths.

[0081] 2. The method for automatic identification of critical paths in urban road networks based on trajectory segmentation described in the present invention proposes screening criteria for starting intersections and calibrates the value range of the criticality of vehicle path decision chains; proposes screening rules for vehicle path decision chains that constitute critical paths and a critical path splicing process, thereby achieving efficient identification of critical paths in urban road networks based on vehicle path decision chain splicing.

[0082] 3. The method for automatic identification of critical paths in urban road networks based on trajectory segmentation described in the present invention first screens the vehicle path decision chains by their criticality before connecting them to form a critical path. Compared with the existing research method of first connecting all local paths to form candidate critical paths and then obtaining the critical path through screening, this method greatly reduces the workload and improves the overall efficiency of critical path identification.

[0083] 4. The trajectory segmentation-based automatic identification method for critical paths in urban road networks, described in this paper, uses criticality as the basis for screening vehicle routing decision chains. This method cleverly avoids a major flaw of existing research methods, which rely on field traffic flow data to calculate average turn ratios, distribute traffic at downstream intersections, and then identify critical paths based on traffic distribution at downstream intersections. This traditional method easily leads to inconsistent path flow estimates with actual conditions, thus affecting the accuracy of critical path estimation and ultimately adversely affecting signal timing design. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] The present invention will be further described below in conjunction with the accompanying drawings:

[0085] Figure 1 This is a flowchart of a method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0086] Figure 2This is a flowchart of a critical path forming method in a method for automatically identifying critical paths in an urban road network based on trajectory segmentation according to the present invention;

[0087] Figure 3 A schematic diagram of the road network structure of the method for automatically identifying key paths in an urban road network based on trajectory segmentation according to the present invention;

[0088] Figure 4 A schematic diagram of a vehicle path decision chain defined by an upstream intersection incoming traffic flow and a downstream intersection outgoing traffic flow in an automatic identification method of an urban road network key path based on trajectory segmentation according to the present invention;

[0089] Figure 5a This is a schematic diagram of the vehicle path decision chain of categories 1, 2, and 3 in the method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0090] Figure 5b This is a schematic diagram of the vehicle path decision chain of categories 4, 5, and 6 in the method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0091] Figure 5c This is a schematic diagram of the vehicle path decision chain of categories 7, 8, and 9 in the method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0092] Figure 6a A schematic diagram of the connection rules between the vehicle path decision chains of categories 1, 2, and 3 and the vehicle path decision chains of categories 1, 4, and 7 in the method for automatically identifying key paths in an urban road network based on trajectory segmentation according to the present invention;

[0093] Figure 6b A schematic diagram of the connection rules between the vehicle path decision chains of categories 4, 5, and 6 and the vehicle path decision chains of categories 2, 5, and 8 in the method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0094] Figure 6c A schematic diagram of the connection rules between the vehicle path decision chains of categories 7, 8, and 9 and the vehicle path decision chains of categories 3, 6, and 9 in the method for automatically identifying key paths in urban road networks based on trajectory segmentation according to the present invention;

[0095] Figure 7 This is a schematic diagram of path composition in a method for automatically identifying key paths in an urban road network based on trajectory segmentation according to the present invention;

[0096] Figure 8This is a schematic diagram of a critical path in an urban road network formed by the automatic identification method of the critical path in an urban road network based on trajectory segmentation described in the present invention.

[0097] In the figure: 1. Renmin Avenue, 2. Jiefang Avenue, 3. Xinmin Square, 4. Ziyou Avenue, 5. Huimin Road, 6. Tsinghua Road, 7. Longli Road, 8. Tongzhi Street, 9. Guilin Road, 10. Xinjiang Street, 11. Dongmin Avenue, 12. Baihui Street, 13. Yongchang Road, 14. Yihe Road, 15. Fengshun Street, 16. Xikang Road, 17. Mingde Road, 18. Lixin Street, 19. Mudan Street, 20. Section l, 21. Incoming traffic flow at upstream intersection i, 22. Outgoing traffic flow at downstream intersection j. DETAILED DESCRIPTION

[0098] The present invention will be described in detail below with reference to the accompanying drawings:

[0099] See Figure 1 The method for automatically identifying key paths in urban road networks based on trajectory segmentation comprises the following steps:

[0100] 1. Obtain the complete continuous driving trajectory of the vehicle based on the checkpoint data

[0101] Read real-time vehicle passing data detected by the checkpoint video detection equipment, including the geographical coordinates of the checkpoint, vehicle arrival time, and vehicle license plate information; intelligently process the vehicle passing data detected by all checkpoint equipment in the area, and classify and integrate the vehicle passing records of the same vehicle at different checkpoints based on the vehicle license plate information to form a complete vehicle passing data set for each vehicle; to ensure that the complete vehicle passing data set obtained is formed by a single trip within a certain time range, the continuity of the checkpoint data is judged and screened; use ArcGIS software to perform road network map matching to obtain the complete and continuous driving trajectory of the vehicle; the steps include the following:

[0102] 1) Collect real-time vehicle passing data detected by the checkpoint video detection equipment:

[0103] x i =(num,p i ,t i )

[0104] Where: i is the bayonet number, x i is the vehicle passing data at checkpoint i, num is the vehicle license plate, p is the coordinates of the vehicle's trajectory point, which is determined according to the position of the checkpoint equipment; t is the time when the vehicle arrives at the checkpoint;

[0105] 2) Through intelligent processing of vehicle passing data detected by all checkpoints in the area, the passing records of the same vehicle at different checkpoints are classified and integrated according to the vehicle license plate information to form a complete set of vehicle passing data for each vehicle:

[0106] If the num in the passing data is the same, it is the passing data of the same vehicle and is integrated into the passing point set X of the vehicle. num In, that is

[0107] X num ={x1,…,x i ,…,x n}

[0108] 3) To avoid interference of vehicle stops on path identification, the continuity of the checkpoint data is judged and screened:

[0109] This step is mainly to ensure that the complete vehicle passing data set is formed by a single trip within a certain time range, so the continuity of the checkpoint data is judged and screened;

[0110]

[0111] Where: t is the time measurement value for judging the continuity of the trajectory; It is the average distance from the stop line in a certain direction of the upstream intersection to the stop line in a certain direction of the downstream intersection of all adjacent intersections in the road network; The average speed of all valid vehicles in the road network, which can be determined based on the historical speed data of the road network area; is the average red light time in all directions of all intersections in the road network; η is the disturbance factor, which is determined according to the specific conditions of the vehicle driving process;

[0112]

[0113] Where: d y,a,b The distance from the stop line in the direction of a on the entrance road of intersection y to the stop line in the direction of b on the entrance road of intersection y+1;

[0114] Δt i =t i+1 -t i

[0115] If Δt i >t, then n=i

[0116] Use the loop structure to verify Δt in sequence i , if Δt i ≤t, then i=i+1, until Δt i >t, the loop ends, and n=i.

[0117] Where: Δt i For a car from trajectory point p i To trajectory point p i+1 The time taken, n is the set of vehicles with license plate number num passing by X num The number of vehicles that have passed the test;

[0118] 4) Use ArcGIS software to perform road network map matching to obtain the complete continuous driving trajectory of the vehicle:

[0119] Using ArcGIS software, the passing points of each vehicle are set as X num ={x1,…,x i ,…,x n Matching with the map of the road network under study to form a complete and continuous driving trajectory of each vehicle in real time;

[0120] 2. Form a vehicle path decision chain, encode it and calculate traffic flow statistics

[0121] The complete vehicle trajectory is segmented according to the segmentation principle to form several vehicle routing decision chains. Then, the vehicle routing decision chains are encoded according to the classification and coding principles, and the traffic flow on the vehicle routing decision chains is counted to provide data support for subsequent traffic analysis and route screening. The steps include the following:

[0122] 2) Segmentation principle:

[0123] See Figure 4 , taking the upstream intersection inflow and downstream intersection outflow between two intersections as a minimum segmentation unit, the complete vehicle driving trajectory is combined according to the upstream arrival and downstream outflow traffic flows of two adjacent intersections to form a vehicle path decision chain;

[0124] 2) Classification principles:

[0125] See Figure 5a 、 5b Similar to 5c, the combination of the upstream arrival traffic flow and the downstream departure traffic flow of vehicles at the intersection is used as the basis for classification, that is,

[0126] l=(i,j)

[0127] vpdc=(l,m,n)

[0128] Where: l represents the road section; i, j represent the upstream and downstream intersection numbers of road section l respectively; m represents the upstream arrival traffic flow attribute of the intersection, m∈{ZX,ZZ,YZ}; n represents the downstream departure traffic flow attribute of the intersection, n∈{ZX,ZZ,YZ};

[0129] According to the combination of m and n, the vehicle routing decision chain can be divided into 9 categories, as shown in the following table:

[0130] Table 1 Vehicle routing decision chain categories

[0131]

[0132] Explanation of symbols in Table 1: vpdc t represents the vehicle routing decision chain of category t, where t∈{1,2,…,9}; ZX represents the through traffic flow; ZZ represents the left-turn traffic flow; and YZ represents the right-turn traffic flow.

[0133] See Figure 5a 、 5b ,5c, the vehicle routing decision chain categories shown in the figure correspond to the vehicle routing decision chain categories in Table 1, see Figure 5a , the types of vehicle routing decision chains are explained. When the incoming traffic flow attribute of the upstream intersection is straight, it can be combined with the straight, left-turn and right-turn traffic flows of the downstream intersection to form three types of vehicle routing decision chains, namely vpdc1: the incoming traffic flow attribute of the upstream intersection is straight, and the outgoing traffic flow attribute of the downstream intersection is straight; vpdc2: the incoming traffic flow attribute of the upstream intersection is straight, and the outgoing traffic flow attribute of the downstream intersection is left-turn; vpdc3: the incoming traffic flow attribute of the upstream intersection is straight, and the outgoing traffic flow attribute of the downstream intersection is right-turn.

[0134] 3) Coding principles:

[0135] The vehicle routing decision chain code should include vehicle trajectory information, intersection number information and vehicle routing decision chain category information, that is,

[0136] v k =(k,l,vpdc t )

[0137] Where: v k represents the kth vehicle path decision chain from top to bottom into which the complete vehicle driving trajectory is divided; l indicates that the kth vehicle path decision chain belongs to road section l; vpdc t represents that the kth vehicle routing decision chain belongs to category t in the vehicle routing decision chain classification, where t∈{1,2,…,9};

[0138] 4) Statistics of each v i Flow value:

[0139] If the code v i If l and t are equal at the same time, they are considered to be the same and are recorded as The flow rate value is recorded as q l,t .

[0140] 3. Based on the connection rules and criticality calculation formula of the vehicle path decision chain, the critical path in the road network is formed

[0141] See Figure 2 、 3 8. Based on the definition and classification principle of vehicle path decision chains, the connection rules of vehicle path decision chains are given; based on the purpose of the invention, a criticality calculation formula of the vehicle path decision chain is creatively proposed, and the value range of the criticality is determined. Subsequently, the starting intersection is screened and the criticality of all vehicle path decision chains in the road network is calculated using the criticality calculation formula. The vehicle path decision chains that constitute the critical path are gradually screened and the vehicle path decision chains that meet the requirements are connected based on the connection rules; all connection paths are output, and the criticality of the vehicle path decision chains in each connection path is accumulated. The connection paths whose accumulated sum ranks in the top 60% are taken as the critical paths in the road network and are placed in the critical path set CP = {p n}, including the following steps:

[0142] 1) Connection rules between vehicle routing decision chains:

[0143] See Figure 6a 、 6b As with 6c, the definition and classification principles of vehicle routing decision chains show that vehicle routing decision chains cannot be connected arbitrarily, but must follow certain rules. Figure 6a ,Since the departure actions of the vehicles vpdc1, vpdc4, and vpdc7 in the figure are all straight, the arrival actions of the vehicle routing decision chains connected to them downstream can only be straight. This means that the vehicle routing decision chain with the combination type of vpdc1, vpdc4, and vpdc7 can only be connected to the vehicle routing decision chain with the combination type of vpdc1, vpdc2, and vpdc3;

[0144] See Figure 6b and Figure 6c Based on the same principle, the connection rules of the vehicle routing decision chain are as follows:

[0145] (4) The vehicle routing decision chains of combination type vpdc1, vpdc4, and vpdc7 can only be connected to the vehicle routing decision chains of combination type vpdc1, vpdc2, and vpdc3;

[0146] (5) The vehicle routing decision chains with the combination type of vpdc2, vpdc5, and vpdc8 can only be connected to the vehicle routing decision chains with the combination type of vpdc4, vpdc5, and vpdc6;

[0147] (6) The vehicle routing decision chains of combination type vpdc3, vpdc6, and vpdc9 can only be connected to the vehicle routing decision chains of combination type vpdc7, vpdc8, and vpdc9;

[0148] 2) Propose a formula for calculating the criticality of the vehicle routing decision chain:

[0149] See Figure 4 , the criticality calculation formula of the vehicle routing decision chain is:

[0150]

[0151] Where: represents the criticality of the vehicle routing decision chain of category t on road segment l 20; l,t represents the flow of the vehicle routing decision chain of category t on road segment 20; q l,t,m represents the total flow of upstream inflow traffic flow m of the vehicle routing decision chain of category t on road segment l 20; l,t,n represents the total flow of downstream outgoing traffic flow n of the vehicle routing decision chain of category t on road segment l 20; represents the total traffic flow in the direction of the d1 entrance road at intersection i on road segment l 20, where d1∈{N,S,W,E}; Represents the total traffic flow in the direction of the d2 entrance road at intersection j on section l 20, where d2∈{N,S,W,E}.

[0152] 3) Determine the value range of criticality:

[0153] The first term in the VDRC criticality calculation formula represents the percentage of the flow of VDRC type t on road segment l 20 to the total flow of the three types of VDRCs formed by the same upstream inflow traffic flow.

[0154] The second term represents the percentage of the upstream incoming traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total flow of the entrance road;

[0155] The third term represents the percentage of the downstream outflow traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total traffic flow of the entrance road;

[0156] It is not difficult to find that when all three percentages in the formula are 1, the criticality value reaches the maximum value of 3, which means that there is only one type of vehicle path decision chain on road section 1 20. The purpose of the automatic identification method of urban road network critical paths based on trajectory segmentation described in this invention is to identify critical paths in urban road networks. Based on expert experience and experimental verification results, the lower limit of the criticality value range is determined to be 1.2. In summary, the criticality value range is [1.2, 3].

[0157] 4) Filter the starting intersection and calculate the criticality of all vehicle routing decision chains in the road network:

[0158] Determine the road network to be studied, use the total flow rate at the intersection as the basis for judging the importance of the intersection, sort the intersections in the road network from most important to least important, and select the intersection with the largest flow rate as the starting intersection. Once an intersection has been used as the starting intersection, the intersection within the six adjacent intersections on the same trunk line cannot be used as the starting intersection. From these intersections, the intersection with the largest flow rate must be selected as the starting intersection. Then, the criticality of all vehicle routing decision chains in the road network is calculated using the criticality calculation formula for vehicle routing decision chains.

[0159] 5) Screening and connection of vehicle routing decision chains that constitute the connection path:

[0160] See Figure 7 After determining the starting intersection, find the two vehicle routing decision chains with the largest criticality in the starting intersection; according to the connection rule between vehicle routing decision chains, find all vehicle routing decision chains that can be connected to the two vehicle routing decision chains with the largest criticality, and judge the criticality of all vehicle routing decision chains that meet the requirements. If the criticality of a vehicle routing decision chain is within the interval [1.2,3], connect the vehicle routing decision chain with the upstream vehicle routing decision chain; then find all vehicle routing decision chains that can be connected to the vehicle routing decision chain at the downstream intersection of the vehicle routing decision chain that has just been connected, and judge the criticality of all vehicle routing decision chains that meet the requirements. If the criticality of a vehicle routing decision chain is within the interval [1.2,3], connect the vehicle routing decision chain with the upstream vehicle routing decision chain; repeat this process until one of the following conditions is met:

[0161] (1) The criticality of all connected vehicle routing decision chains is not within the interval [1.2,3];

[0162] (2) When the number of vehicle path decision chains in the connected path is 6;

[0163] (3) There are already two vehicle path decision chains that meet the requirements and are connected at the intersection;

[0164] Based on the design and effectiveness of signal coordination control schemes in actual urban road networks, we found that when implementing signal coordination control, the number of coordinated intersections cannot be increased indefinitely. If the number is too large, the coordination effect will be greatly reduced. Generally, the number of coordinated intersections should not exceed 6. Therefore, the number of vehicle path decision chains included in the formed critical path should not exceed 5.

[0165] Repeat the above process until there is no intersection in the road network that can be used as the starting intersection.

[0166] 6) Forming the critical path in the road network:

[0167] The criticality of the vehicle path decision chain in each connection path is accumulated, and the connection paths with the top 60% of the accumulated sum are taken as the critical paths in the road network and put into the critical path set CP = {p n}middle.

[0168] Example:

[0169] See Figure 3 、 6a , 6b, 6c and 7, in order to verify the effectiveness of the automatic identification method of key paths in urban road networks based on trajectory segmentation, an urban road network was built based on Vissim simulation software and the vehicle driving paths and traffic were input to obtain the complete driving trajectory of each vehicle. The vehicle driving trajectories were segmented and encoded according to the segmentation principle and the coding principle, and the vpdc k,t The flow value q k,t Then, the total traffic volume at the intersection is used as the basis for judging the importance of the intersection. The intersections in the road network are sorted from most important to least important, and the intersection with the largest traffic volume is selected as the starting intersection. Once an intersection has been used as the starting intersection, the intersections within the six adjacent intersections on the same trunk line cannot be used as the starting intersection. The intersection with the largest traffic volume needs to be selected from the remaining intersections as the starting intersection. The criticality calculation formula of the vehicle routing decision chain proposed in the present invention is used to calculate the criticality of all vehicle routing decision chains in the road network, and the two vehicle routing decision chains with the largest criticality at the starting intersection are found. According to the connection rule between vehicle routing decision chains, all vehicle routing decision chains that can be connected to the two vehicle routing decision chains with the largest criticality are found respectively, and the criticality of all vehicle routing decision chains that meet the requirements is judged. If the criticality of a vehicle routing decision chain is within the interval [1.2,3], the vehicle routing decision chain is connected to the upstream vehicle routing decision chain. Then, all vehicle routing decision chains that can be connected to the vehicle routing decision chain that has just been connected are found at the downstream intersection, and the criticality of all vehicle routing decision chains that meet the requirements is judged. If the criticality of a vehicle routing decision chain is within the interval [1.2,3], the vehicle routing decision chain is connected to the upstream vehicle routing decision chain. This process is repeated until one of the following conditions is met:

[0170] 1. The criticality of all connected vehicle routing decision chains is not within the interval [1.2,3];

[0171] 2. When the number of vehicle routing decision chains in the connected path is 6;

[0172] 3. There are already two vehicle path decision chains that meet the requirements and are connected at the intersection.

[0173] Repeat the above process until there is no intersection in the road network that can be used as the starting intersection. Finally, output all the connection paths and accumulate the criticality of the vehicle path decision chain in each connection path. Take the connection paths that rank in the top 60% of the cumulative sum as the critical paths in the road network and put them into the critical path set CP = {p n}middle.

[0174] The specific composition of each critical path output by the simulation is shown in the following table:

[0175] Table 2 Specific composition of each critical path

[0176]

[0177] Note: The table only gives the starting and ending points of the critical path. The specific composition of the critical path needs to be combined with the table and Figure 8 The key path diagram of the urban road network is determined.

[0178] The critical paths in the road network formed by the automatic identification method of the urban road network critical paths based on trajectory segmentation described in the present invention are basically the same as the critical paths in the actual road network, and the effectiveness of the method of the present invention is verified.

[0179] Taking critical path 2 as an example, the formation process of the critical path is introduced in detail.

[0180] The criticality calculation formula of the vehicle path decision chain The criticality of all vehicle routing decision chains in the road network is calculated. According to the definition rules of the starting intersection, the intersection of Baihui Street and Jiefang Road is selected as the starting intersection. The top two vehicle routing decision chains at the intersection of Baihui Street and Jiefang Road are and in The critical degree of is not in [1.2,3] and is therefore discarded, so only Then to The criticality of the vehicle routing decision chain that satisfies the connection rules between the vehicle routing decision chains at the downstream is judged, and it is found that only The critical degree of is in [1.2,3], so and Connect and repeat the above process until the number of vehicle routing decision chains in the connected path reaches 5. At this time, the last vehicle routing decision chain is The starting and ending points of the resulting critical path are the east intersection of Baihui Street and Jiefang Road, and the north intersection of Dongmin Street and Jiefang Road, respectively. The critical path connection is terminated because the number of coordinated intersections cannot be increased indefinitely during signal coordination. Excessive coordination significantly reduces the coordination effect. Generally, the number of coordinated intersections should not exceed six. Therefore, the number of vehicle routing decision chains included in the resulting critical path should not exceed five.

[0181] The above experimental process proves that the method for automatic identification of critical paths in urban road networks based on trajectory segmentation described in the present invention not only overcomes the problem of insufficient consideration of dynamic traffic flow conditions in existing node-based road network critical path identification methods while completing automatic identification of critical paths, but also overcomes the problem that the road network critical path identification methods based on key sections have large differences in corresponding theoretical models and evaluation indicators due to different research perspectives, resulting in large discrepancies between research results and actual road network conditions, that is, the credibility of the formed critical paths is not high.

Claims

1. A method for automatically identifying key paths in urban road networks based on trajectory segmentation, characterized by The method for automatically identifying key paths in urban road networks based on trajectory segmentation includes the following steps: 1) Obtain the complete and continuous driving trajectory of the vehicle based on the checkpoint data; 2) Form a vehicle path decision chain and encode it and calculate traffic flow statistics; 3) Based on the connection rules and criticality calculation formula of the vehicle path decision chain, the critical path in the road network is formed.

2. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 1 is characterized in that ,The steps of obtaining the complete continuous driving trajectory of the vehicle based on the ,checkpoint data are as follows: 1) Collect real-time vehicle passing data detected by the checkpoint video detection equipment: x i =(num,p i ,t i ) Where: i is the bayonet number, x i is the vehicle passing data at checkpoint i, num is the vehicle license plate, p is the coordinates of the vehicle's trajectory point, which is determined according to the position of the checkpoint equipment; t is the time when the vehicle arrives at the checkpoint; 2) Through intelligent processing of vehicle passing data detected by all checkpoints in the area, the passing records of the same vehicle at different checkpoints are classified and integrated according to the vehicle license plate information to form a complete set of vehicle passing data for each vehicle: If the num in the passing data is the same, it is the passing data of the same vehicle and is integrated into the passing point set X of the vehicle. num In, that is X num ={x1,…,x i ,…,x n } 3) To avoid interference of vehicle stops on path identification, the continuity of the checkpoint data is judged and screened: This step is mainly to ensure that the complete vehicle passing data set is formed by a single trip within a certain time range, so the continuity of the checkpoint data is judged and screened; Where: t is the time measurement value for judging the continuity of the trajectory; It is the average distance from the stop line in a certain direction of the upstream intersection to the stop line in a certain direction of the downstream intersection of all adjacent intersections in the road network; The average speed of all valid vehicles in the road network, which can be determined based on the historical speed data of the road network area; is the average red light time in all directions of all intersections in the road network; η is the disturbance factor, which is determined according to the specific conditions of the vehicle driving process; Where: d y,a,b The distance from the stop line in the direction of a on the entrance road of intersection y to the stop line in the direction of b on the entrance road of intersection y+1; Δt i =t i+1 -t i If Δt i >t, then n=i Use the loop structure to verify Δt in sequence i , if Δt i ≤t, then i=i+1, until Δt i >t, the loop ends, and n=i. Where: Δt i For a car from trajectory point p i To trajectory point p i+1 The time taken, n is the set of vehicles with license plate number num passing by X num The number of vehicles that have passed the test; 4) Use ArcGIS software to perform road network map matching to obtain the complete continuous driving trajectory of the vehicle: Using ArcGIS software, the passing points of each vehicle are set as X num ={x1,…,x i ,…,x n } is matched with the map of the road network under study to form a complete and continuous driving trajectory of each vehicle in real time.

3. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 1 is characterized in that ,The steps of forming the vehicle path decision chain, encoding it and traffic flow statistics are as follows: 1) Segmentation principle: Taking the upstream intersection inflow and downstream intersection outflow between two intersections as a minimum segmentation unit, the complete vehicle driving trajectory is combined according to the upstream arrival and downstream outflow traffic flows of two adjacent intersections to form a vehicle path decision chain; 2) Classification principles: The combination of the upstream arrival traffic flow and the downstream departure traffic flow of vehicles at the intersection is used as the basis for classification, that is, l=(i,j) vpdc=(l,m,n) Where: l represents the road section; i, j represent the upstream and downstream intersection numbers of road section l respectively; m represents the upstream arrival traffic flow attribute of the intersection, m∈{ZX,ZZ,YZ}; n represents the downstream departure traffic flow attribute of the intersection, n∈{ZX,ZZ,YZ}; According to the combination of m and n, the vehicle routing decision chain can be divided into 9 categories, as shown in the following table: Table 1 Vehicle routing decision chain categories Explanation of symbols in Table 1: vpdc t represents the vehicle routing decision chain of category t, where t∈{1,2,…,9}; ZX represents the through traffic flow; ZZ represents the left-turn traffic flow; and YZ represents the right-turn traffic flow. The vehicle routing decision chain categories shown in Figures 5a, 5b, and 5c correspond to the vehicle routing decision chain categories in Table 1. The types of vehicle routing decision chains are explained. When the inflow traffic flow attribute of the upstream intersection is straight, it can be combined with the straight, left-turn, and right-turn traffic flows of the downstream intersection to form three types of vehicle routing decision chains, namely, vpdc1: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is straight; vpdc2: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is left-turn; vpdc3: the inflow traffic flow attribute of the upstream intersection is straight, and the outflow traffic flow attribute of the downstream intersection is right-turn; 3) Coding principles: The vehicle routing decision chain code should include vehicle trajectory information, intersection number information and vehicle routing decision chain category information, that is, v k =(k,l,vpdc t ) Where: v k represents the kth vehicle path decision chain from top to bottom into which the complete vehicle driving trajectory is divided; l indicates that the kth vehicle path decision chain belongs to road section l; vpdc t represents that the kth vehicle routing decision chain belongs to category t in the vehicle routing decision chain classification, where t∈{1,2,…,9}; 4) Statistics of each v i Flow value: If the code v i If l and t are equal at the same time, they are considered to be the same and are recorded as The flow rate value is recorded as q l,t .

4. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 1 is characterized in that ,Based on the connection rules and criticality calculation formula of the vehicle path decision chain, the steps of forming the critical path in the road network are as follows: 1) Provide the connection rules between vehicle routing decision chains; 2) Propose a formula for calculating the criticality of the vehicle routing decision chain; 3) Determine the value range of criticality; 4) Screen the starting intersection and calculate the criticality of all vehicle routing decision chains in the road network; 5) Screening and connecting vehicle routing decision chains that constitute the connection path; 6) Form the critical path in the road network.

5. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 4 is characterized in that ,The connection rules between the vehicle path decision chains mentioned are: The definition and classification principles of vehicle routing decision chains show that vehicle routing decision chains cannot be connected arbitrarily, but must follow certain rules. Since the vehicle departure actions of vpdc1, vpdc4, and vpdc7 in Figure 6a are all straight, the arrival actions of the vehicle routing decision chains connected to them downstream can only be straight. This means that the vehicle routing decision chain with the combination type of vpdc1, vpdc4, and vpdc7 can only be connected to the vehicle routing decision chain with the combination type of vpdc1, vpdc2, and vpdc3. Based on the same principle, the connection rules of the vehicle routing decision chain are as follows: (1) The VDRs of combination type vpdc1, vpdc4, and vpdc7 can only be connected to the VDRs of combination type vpdc1, vpdc2, and vpdc3; (2) The vehicle routing decision chains with the combination type of vpdc2, vpdc5, and vpdc8 can only be connected to the vehicle routing decision chains with the combination type of vpdc4, vpdc5, and vpdc6; (3) The vehicle routing decision chains of combination type vpdc3, vpdc6, and vpdc9 can only be connected to the vehicle routing decision chains of combination type vpdc7, vpdc8, and vpdc9.

6. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 4 is characterized in that ,The criticality calculation formula of the proposed vehicle path decision chain is : Where: represents the criticality of the vehicle routing decision chain of category t on road segment l 20; l,t represents the flow of the vehicle routing decision chain of category t on road segment 20; q l,t,m represents the total flow of upstream inflow traffic flow m of the vehicle routing decision chain of category t on road segment l 20; l,t,n represents the total flow of downstream outgoing traffic flow n of the vehicle routing decision chain of category t on road segment l 20; represents the total traffic flow in the direction of the d1 entrance road at intersection i on road segment l 20, where d1∈{N,S,W,E}; Represents the total traffic flow in the direction of the d2 entrance road at intersection j on section l 20, where d2∈{N,S,W,E}.

7. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 4 is characterized in that ,The value range of determining the criticality is: The first term in the VDRC criticality calculation formula represents the percentage of the flow of VDRC type t on road segment l 20 to the total flow of the three types of VDRCs formed by the same upstream inflow traffic flow. The second term represents the percentage of the upstream incoming traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total flow of the entrance road; The third term represents the percentage of the downstream outflow traffic flow of the vehicle routing decision chain of type t on road segment l 20 to the total traffic flow of the entrance road; It is not difficult to find that when the three percentages in the formula are all 1, the criticality value reaches the maximum value of 3, which means that there is only one type of vehicle path decision chain on the road section 20. The purpose of the automatic identification method of urban road network critical paths based on trajectory segmentation described in the present invention is to identify critical paths in urban road networks. According to expert experience and experimental verification results, the lower limit of the criticality value range is determined to be 1.2; in summary, the criticality value range is [1.2, 3].

8. According to the method of automatic identification of key paths in urban road networks based on trajectory segmentation as claimed in claim 4, it is characterized in that ,The ,said screening starting intersection and calculating the criticality of all vehicle path decision chains in the road network ,refers to: Determine the road network to be studied, use the total flow rate of the intersection as the basis for judging the importance of the intersection, sort the intersections in the road network from large to small according to their importance, and select the intersection with the largest flow rate as the starting intersection. After an intersection has been used as the starting intersection, the intersections within the six adjacent intersections on the same trunk line cannot be used as the starting intersection. It is necessary to select the intersection with the largest flow rate from these intersections as the starting intersection; then, calculate the criticality of all vehicle path decision chains in the road network using the criticality calculation formula of the vehicle path decision chain.

9. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 4 is characterized in that ,The screening and connection of the vehicle path decision chains constituting the ,connection path refers to: After determining the starting intersection, find the two VDRs with the highest criticality at the starting intersection. According to the connection rules between VDRs, find all VDRs that can be connected to the two VDRs with the highest criticality. Determine the criticality of all VDRs that meet the requirements. If the criticality of a VDR is within the interval [1.2,3], connect the VDR to the upstream VDR. Then, find all VDRs that can be connected to the VDR at the downstream intersection of the VDR that has just been connected. Determine the criticality of all VDRs that meet the requirements. If the criticality of a VDR is within the interval [1.2,3], connect the VDR to the upstream VDR. Repeat this process until one of the following conditions is met: (1) The criticality of all connected vehicle routing decision chains is not within the interval [1.2,3]; (2) When the number of vehicle path decision chains in the connected path is 6; (3) There are already two vehicle path decision chains that meet the requirements and are connected at the intersection; Repeat the above process until there is no intersection in the road network that can be used as the starting intersection.

10. The method for automatically identifying key paths in urban road networks based on trajectory segmentation according to claim 4 is characterized in that ,The critical path in the road network mentioned refers to: Output all connection paths, and accumulate the criticality of the vehicle path decision chain in each connection path, take the connection paths that rank in the top 60% of the cumulative sum as the critical paths in the road network and put them into the critical path set CP = {p n }middle.

Images