Trunk line scene recognition method and device and electronic equipment

Abstract

The invention provides a trunk line scene recognition method and device and electronic equipment. The method comprises the following steps: acquiring static road network information and dynamic traffic flow information of a to-be-identified trunk line and a trunk line associated road section in a target time period; according to the static road network information and the dynamic traffic flow information of the to-be-identified trunk line and the trunk line associated road section in the target time period, extracting a plurality of traffic flow characteristics of the to-be-identified trunk line; and comparing the traffic flow characteristics of the to-be-identified trunk line with corresponding traffic flow characteristics of a trunk line prototype in a trunk line prototype scene library, and determining a scene category of the to-be-identified trunk line according to a comparison result. According to the invention, multiple traffic flow characteristics of the to-be-identified trunk line can be obtained according to the static road network information and the dynamic traffic flow information of the to-be-identified trunk line and the trunk line associated road section, the scene category of the to-be-identified trunk line is determined more finely, and a good decision basis is provided for trunk line coordination control.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic control optimization, and in particular to a trunk line scene recognition method, device and electronic equipment. Background Art

[0002] Arterial coordinated control is an important technical means to optimize the efficiency of traffic operation in the main flow direction of the main sections of the city. Different arterial coordinated control strategies are potentially applicable to different spatiotemporal characteristics of arterial traffic flow. In order to improve the efficiency of traffic operation, it is necessary to identify arterial scenes so as to provide corresponding arterial coordinated control strategies based on the identified arterial traffic flow status. Arterial scene recognition is to identify the arterial traffic flow status that has a potential corresponding relationship with the arterial coordinated control strategy by analyzing the traffic flow characteristics of the urban road arterial in a specific direction and time period, and provide decision support for the optimization selection of the arterial coordinated control scheme.

[0003] At present, traffic state recognition technology is usually used when conducting coordinated control of trunk lines. Based on macro traffic parameters such as flow, speed, occupancy, saturation, etc., threshold division, fuzzy evaluation, clustering, classification and other methods are used to identify frequent or occasional congestion and its degree, thereby distinguishing the trunk lines and providing services for control scheme selection, traffic induction, accident handling, etc.

[0004] However, the existing traffic status recognition technology can only evaluate the congestion level of traffic flow in a single dimension, and the recognition features used are too macro and not sufficiently consistent with the traffic characteristics of different scenarios such as urban intersections, trunk roads, and regional road networks. Therefore, it cannot fully meet the actual needs of trunk road scene recognition. Summary of the invention

[0005] The trunk line scene recognition method, device, electronic device and computer-readable storage medium provided by the present invention can obtain multiple traffic flow characteristics of the trunk line to be identified based on the static road network information of the trunk line to be identified and the trunk-related sections and the dynamic traffic flow information, and more finely determine the scene category of the trunk line to be identified, thereby providing a good decision-making basis for trunk line coordinated control.

[0006] In a first aspect, the present invention provides a trunk scene recognition method, the method comprising:

[0007] Obtain static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line-related sections during the target period;

[0008] Extracting multiple traffic flow features of the trunk line to be identified based on the static road network information and dynamic traffic flow information of the trunk line to be identified and the road sections associated with the trunk line during the target period;

[0009] The traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of the trunk line prototypes in the trunk line prototype scene library, and the scene category of the trunk line to be identified is determined according to the comparison result.

[0010] Optionally, before acquiring the static road network information and dynamic traffic flow information of the trunk to be identified and the trunk-associated road sections in the target time period, the method further includes:

[0011] A trunk line prototype scenario library is constructed, wherein the trunk line prototype scenario library includes a plurality of red wave scenario trunk line prototypes and a plurality of non-red wave scenario trunk line prototypes, and each red wave scenario trunk line prototype and each non-red wave scenario trunk line prototype correspond to a plurality of traffic flow features respectively.

[0012] Optionally, comparing the traffic flow characteristics of the trunk line to be identified with corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determining the scene category of the trunk line to be identified according to the comparison result includes:

[0013] The traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of all red wave scenario trunk line prototypes and all non-red wave scenario trunk line prototypes in the trunk line prototype scenario library, and whether the trunk line to be identified is a red wave scenario trunk line is determined according to the comparison results.

[0014] Optionally, comparing the traffic flow characteristics of the trunk line to be identified with corresponding traffic flow characteristics of all red wave scene trunk line prototypes and all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and determining whether the trunk line to be identified is a red wave scene trunk line according to the comparison result includes:

[0015] For each traffic flow feature required for red wave scene recognition, obtain the average difference between the trunk line to be recognized and all trunk line prototypes of red wave scenes under the same traffic flow feature, and the average difference between the trunk line to be recognized and all trunk line prototypes of non-red wave scenes under the same traffic flow feature;

[0016] If the average difference between the trunk line to be identified and all trunk line prototypes in red wave scenarios under the same traffic flow feature is less than the average difference between the trunk line to be identified and all trunk line prototypes in non-red wave scenarios under the same traffic flow feature, the traffic flow feature is determined to be a red wave feature; otherwise, the traffic flow feature is determined to be a non-red wave feature;

[0017] Count the number of red wave features and non-red wave features respectively;

[0018] If the number of red wave features is greater than the number of non-red wave features, the trunk line to be identified is determined to be a red wave scene trunk line; otherwise, the trunk line to be identified is determined to be a non-red wave scene trunk line.

[0019] Optionally, the non-red wave scene trunk prototype is divided into multiple dimensions, and each dimension is divided into two different non-red wave scene trunk prototypes;

[0020] If it is determined that the trunk line to be identified is a non-red wave scene trunk line, then comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determining the scene category of the trunk line to be identified according to the comparison result also includes:

[0021] For each dimension, the traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and the non-red wave scene category of the trunk line to be identified under the dimension is determined according to the comparison result.

[0022] Optionally, comparing the traffic flow characteristics of the trunk line to be identified with corresponding traffic flow characteristics of all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and determining the non-red wave scene category of the trunk line to be identified under the dimension according to the comparison result includes:

[0023] Respectively obtain the average difference between the trunk line to be identified and all the first non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension, and the average difference between the trunk line to be identified and all the second non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension;

[0024] If the average difference between the trunk line to be identified and all the first non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension is less than the average difference between the trunk line to be identified and all the second non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension, then the trunk line to be identified is determined to be the first non-red wave scenario trunk line under the dimension; otherwise, the trunk line to be identified is determined to be the second non-red wave scenario trunk line under the dimension.

[0025] Optionally, the static road network information and dynamic traffic flow information of the trunk section to be identified and the trunk-associated section in the target time period include at least one of the following information: section length, direction, free flow speed, upstream and downstream topological relationship, flow rate in each direction, saturation in each direction, green light time of each lane in the section, and vehicle passing records.

[0026] Optionally, after comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library and determining the scene category of the trunk line to be identified according to the comparison result, the method further includes:

[0027] According to the determined scene category of the trunk line to be identified, a corresponding coordination control strategy is selected, specifically including: if the scene category of the trunk line to be identified is a red wave scene, then a red wave coordination control strategy is selected; if the scene category of the trunk line to be identified is a non-red wave scene, then a congestion coordination control strategy or a green wave coordination control strategy is selected according to the non-red wave scene category of the trunk line to be identified in each dimension, wherein the non-red wave scene trunk prototype is divided into multiple dimensions.

[0028] In a second aspect, the present invention provides a trunk scene recognition device, the device comprising:

[0029] An acquisition unit, used to acquire static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated road sections during the target period;

[0030] An extraction unit, configured to extract a plurality of traffic flow features of the trunk line to be identified according to the static road network information and dynamic traffic flow information of the trunk line to be identified and the road sections associated with the trunk line during the target period;

[0031] The identification unit is used to compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototypes in the trunk line prototype scene library, and determine the scene category of the trunk line to be identified according to the comparison result.

[0032] In a third aspect, the present invention provides an electronic device, the electronic device comprising:

[0033] at least one processor; and

[0034] a memory communicatively connected to the at least one processor; wherein,

[0035] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above-mentioned trunk scene recognition method.

[0036] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the above-mentioned trunk scene recognition method is implemented.

[0037] The trunk scene recognition method, device, electronic device and computer-readable storage medium provided in the embodiments of the present invention obtain multiple traffic flow characteristics associated with the trunk to be identified and the trunk coordination control scene based on the static road network information and dynamic traffic flow information of the trunk to be identified and the trunk-related sections, and determine the scene category of the trunk to be identified by comparing the feature differences between the trunk to be identified and multiple different trunk prototypes in the trunk prototype scene library, so that the scene category of the trunk to be identified can be determined more finely, providing a good decision-making basis for trunk coordination control. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a flow chart of a trunk line scene recognition method according to an embodiment of the present invention;

[0039] Figure 2 A flowchart of a trunk scene recognition method according to another embodiment of the present invention;

[0040] Figure 3 A schematic diagram of a trunk line section, a trunk line intersection and associated road sections provided in an embodiment of the present invention;

[0041] Figure 4 A flow chart of calculating the flow rate of the downstream coordinated direction of a trunk line section provided by an embodiment of the present invention;

[0042] Figure 5 A flow chart for calculating the number of periodic stops of a coordinated fleet on a trunk line provided by an embodiment of the present invention;

[0043] Figure 6 A flowchart of red wave scene recognition provided by an embodiment of the present invention;

[0044] Figure 7 A schematic diagram of measuring the difference between the characteristics of two trunk lines provided by an embodiment of the present invention;

[0045] Figure 8 It is a structural schematic diagram of a trunk scene recognition device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] The embodiment of the present invention provides a trunk line scene recognition method, such as Figure 1 As shown, the method includes:

[0048] S11. Obtain static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line-related sections during the target period.

[0049] S12. Extract multiple traffic flow features of the trunk line to be identified based on the static road network information and dynamic traffic flow information of the trunk line to be identified and the road sections associated with the trunk line during the target period.

[0050] S13, comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determining the scene category of the trunk line to be identified according to the comparison result.

[0051] The trunk line scene recognition method provided in the embodiment of the present invention obtains multiple traffic flow characteristics associated with the trunk line to be identified and the trunk line coordinated control scene based on the static road network information of the trunk line to be identified and the trunk line associated road sections and the dynamic traffic flow information, and determines the scene category of the trunk line to be identified by comparing the feature differences between the trunk line to be identified and multiple different trunk line prototypes in the trunk line prototype scene library, thereby being able to more finely determine the scene category of the trunk line to be identified, providing a good decision-making basis for trunk line coordinated control.

[0052] The trunk line scene recognition method of the present invention is described in detail below with reference to specific embodiments.

[0053] like Figure 2 As shown, this embodiment specifically includes the following steps:

[0054] Step S21, constructing a trunk prototype scenario library.

[0055] The trunk line prototype scenario library includes a plurality of red wave scenario trunk line prototypes and a plurality of non-red wave scenario trunk line prototypes, and each red wave scenario trunk line prototype and each non-red wave scenario trunk line prototype correspond to a plurality of traffic flow features respectively.

[0056] Furthermore, the trunk line prototypes in non-red wave scenarios can be divided from three dimensions: queuing degree, fleet discreteness, and external interference: from the perspective of queuing degree, they can be divided into slight queuing trunk line prototypes and continuous queuing trunk line prototypes; from the perspective of fleet discreteness, they can be divided into fleet discrete trunk line prototypes and fleet tight trunk line prototypes; from the perspective of external interference, they can be divided into external interference small trunk line prototypes and external interference large trunk line prototypes. There are multiple trunk line prototypes in different dimensions, and each trunk line prototype corresponds to multiple traffic flow characteristics.

[0057] Step S22: Obtain static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line-associated road sections during the target period.

[0058] Specifically, the static road network information and dynamic traffic flow information of the trunk road sections to be identified and the trunk-associated road sections during the target time period may include but are not limited to at least one of the following information: section length, direction, free flow speed, upstream and downstream topological relationship, flow rate in each direction, saturation in each direction, green light time of each lane in the section and vehicle passing records.

[0059] The number of main road intersections to be identified is set to N (N≥2), and the target sections for collecting data for main road scene identification include main road sections, sections whose downstream intersections are N-1 intersections before the main road, and sections whose upstream intersections are the Nth intersection of the main road. In addition, since the present application is for scene recognition of main road traffic flow in a single direction, and "sections whose downstream intersections are N-1 intersections before the main road" include sections in the same direction as the main road section and sections opposite to the main road section, the relevant information of the sections opposite to the main road section will not be used when performing main road scene recognition. Therefore, the information of these sections may not be obtained, that is, the sections opposite to the main road section are removed. Figure 3 shown.

[0060] Step S23, extracting multiple traffic flow features of the trunk line to be identified based on the static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated sections during the target period.

[0061] Specifically, the multiple traffic flow characteristics of the trunk line to be identified extracted may include at least two of the following 10: average saturation of the trunk line segment, upstream and downstream variation trends of the saturation of the trunk line segment, upstream and downstream variation trends of the travel speed of the trunk line segment, upstream and downstream variation trends of the travel time of the trunk line segment, frequency distribution of travel speed of the trunk line segment, flow rate in the coordinated direction downstream of the trunk line segment, flow rate in the non-coordinated direction upstream of the trunk line segment, number of periodic stops of the coordinated fleet of the trunk line segment, and degree of discreteness of the coordinated fleet of the trunk line segment.

[0062] The specific calculation method of traffic flow characteristics includes but is not limited to the following examples.

[0063] (1) Average saturation of trunk line sections

[0064] The average saturation of the trunk line section is the mean of the saturation of the coordinated directions of all the constituent sections of the trunk line, calculated by the following formula:

[0065]

[0066] In the formula, is the average saturation of the trunk line section, sat i is the coordination direction saturation of trunk line section i, and n is the number of trunk line sections.

[0067] (2) Trends of upstream and downstream saturation in trunk line sections

[0068] It is calculated by the following formula:

[0069]

[0070] In the formula, sat trend is the upstream and downstream trend index of saturation of trunk line section, sat n is the saturation of the most downstream section of the trunk line, is the average saturation of the trunk line section.

[0071] (3) Trends of travel speed changes upstream and downstream on trunk lines

[0072] It is calculated by the following formula:

[0073]

[0074] Where speed trend It is the upstream and downstream variation trend index of the travel speed of the trunk line section. is the travel speed of the most downstream section of the trunk line, is the average travel speed on the trunk line.

[0075] (4) Trends in travel time upstream and downstream of trunk line sections

[0076] It is calculated by the following formula:

[0077]

[0078] In the formula, time trend It is the upstream and downstream variation trend index of the travel time of the trunk line section. is the travel time of the most downstream section of the trunk line, is the average travel time on the trunk line.

[0079] (5) Frequency distribution of travel speed on trunk line sections

[0080] Specifically, extract the vehicle samples that pass through all sections of the trunk line within the identification period, and count the frequency distribution F of the samples with a travel speed of 0 to 100 km / h at a statistical interval of 5 km / h (a total of 20) speed .

[0081] For example, the total number of vehicle samples passing through all sections of several lines within the identification period is X, and the number of vehicle samples with travel speed intervals of 0-5km / h, 95-10km / h, ..., 95-100km / h are counted respectively. Then the travel speed frequency distribution of the trunk line section is the number of vehicle samples with travel speed intervals of 0-5km / h, 95-10km / h, ..., 95-100km / h divided by the total number of vehicle samples, which is X.

[0082] (6) Flow rate in the downstream coordinated direction of the trunk line section

[0083] For each component section of the trunk line, the downstream coordination direction flow rate within the scene recognition period is counted. The calculation steps are as follows: Figure 4 As shown, the specific steps are as follows:

[0084] Step S41, extracting the target direction lane of the target road section and determining the lane passing weight.

[0085] If the target direction is straight, the straight, straight left, straight right, and straight left and right lanes of the entrance road of the target section are extracted, and the passing weight of the straight lane is assigned to 1, and the passing weight of the other lanes is assigned to α; if the coordination direction is left turn, the left turn, straight left, left and right, and straight left and right lanes of the entrance road of the section are extracted, and the passing weight of the left turn lane is assigned to 1, the passing weight of the straight left lane is assigned to β1, the passing weight of the left and right lanes is assigned to β2, and the passing weight of the straight left and right lanes is assigned to β3; where α is the proportion of straight vehicles in the straight left, straight right, and straight left and right lanes, and β1, β2, and β3 are the proportions of left-turn vehicles in the straight left, left and right, and straight left and right lanes, respectively.

[0086] Step S42: extract the green light time of the target road section in the target direction.

[0087] Specifically, the green light time in the target direction of the target road section in all cycles within the scene recognition period is extracted.

[0088] Step S43, calculating the traffic flow rate within a single lane and single cycle green light time.

[0089] The number of vehicles passing through a single lane within a single green light cycle is counted at 1-second intervals. The calculation method is as follows:

[0090] r g =|{veh|veh pass-time ∈[g-1,g)}|,g=1,2,K,G

[0091] R lane =[r1L r G ]

[0092] In the formula, r g is the number of vehicles passing through a lane in the gth second of a single green light cycle, which is usually 0 or 1. G is the green light duration of the lane in the cycle, and R lane is the lane single cycle green light time flow rate.

[0093] Step S44, calculating the single-cycle green light time flow rate in the target direction.

[0094] The single-cycle green light time flow rate in the target direction is the weighted sum of the lane flow rates in step S43, and is calculated as follows:

[0095] R turn =∑ lane w lane R lane

[0096] In the formula, R turn is the single-cycle green light time flow rate in the target direction, w lane is the lane passing weight defined in step S41, R laneIt is the lane single cycle green light time flow rate calculated in step S43.

[0097] Step S45: Calculate the average green light time flow rate in the target direction during the scene recognition period.

[0098] Calculate the average green light time flow rate of the target direction in all cycles during the scene recognition period. The calculation method is as follows:

[0099]

[0100] In the formula, R interval is the average green light time flow rate in the target direction of the target road section during the scene recognition period, cycle is the set of all cycles in the scene period, and R turn It is the single-cycle green light time flow rate in the target direction calculated in step S44.

[0101] The coordinated flow rate in the downstream direction of the trunk line segment can be written in the following table form:

[0102]

[0103] In the formula, Coordinate directional flow rates for downstream trunk line segments, is the average flow rate in the downstream coordination direction of road section i during the scene recognition period, and its data structure is a one-dimensional array with a length of its maximum green light duration.

[0104] (7) Flow rate in the upstream non-coordinated direction of the trunk line section

[0105] The calculation method of the flow rate in the upstream non-coordinated direction of the trunk line section is the same as that of the above steps S41 to S45, and is expressed as follows:

[0106]

[0107] In the formula, is the non-coordinated flow rate in the upstream direction of the trunk line section, is the average flow rate in the upstream non-coordinated direction of section i during the scene recognition period, and its data structure is a one-dimensional array with a length of its maximum green light duration.

[0108] (8) Coordinate the number of periodic stops of fleets on trunk routes

[0109] Count the average number of stops of the coordinated controlled convoy on the trunk line when it passes through the green light time of each cycle from the upstream and drives to the downstream. The calculation steps are as follows: Figure 5 As shown, the specific steps are as follows:

[0110] Step S51, extracting the green light time in the upstream coordination direction of the road section.

[0111] Extract the green light start and end times of all cycles in the upstream coordination direction of the trunk line section within the scene recognition period.

[0112] Step S52: extract the coordinated vehicle fleet that passes within the green light time of each cycle.

[0113] The fleet that enters from the upstream intersection of the trunk road section in the coordinated direction during the above green light time and leaves from the downstream intersection of the section in the coordinated direction is extracted, and this fleet is defined as the coordinated fleet.

[0114] Step S53: Calculate the average number of stops of the coordinated fleet in a single cycle.

[0115] The number of stops is calculated based on the theoretical travel time and actual travel time of a single sample, and the average number of stops of the coordinated fleet released in a single cycle in the upstream coordinated direction is calculated.

[0116] Step S54: averaging the average number of stops of all periodically coordinated fleets during the scene recognition period.

[0117] Finally, the number of periodic stops of the coordinated fleet on the trunk line is a one-dimensional array structure, which is expressed as follows:

[0118] P=[P1,K,P n ]

[0119] Where P is the number of stops of the coordinated fleet on the trunk line, P i is the average number of stops of all cycle coordinated fleets on road segment i during the identification period.

[0120] (9) Number of periodic stops of the coordinated fleet on the trunk line downstream

[0121] The number of periodic stops of the coordinated fleet downstream of the trunk line is the number of periodic stops of the coordinated fleet at the most downstream section of the trunk line in step S54. n .

[0122] (10) Dispersion degree of coordinated fleet on trunk lines

[0123] The degree of dispersion of coordinated fleets on trunk lines is defined by the variance of the headway time and relative travel speed data points of the coordinated fleet, and is calculated as follows:

[0124]

[0125] Where D i h is the discrete degree of the coordinated fleet on the road section, v With sd v are the headway time and relative travel speed of vehicle v, N i It is the total number of coordinated vehicles passing through the lane with the largest traffic volume on the road segment during the scene identification period.

[0126] The discrete degree of coordinated fleet on trunk line is a one-dimensional array structure, which is recorded as:

[0127] D=[D1,D2,K,D n ]

[0128] Where D is the degree of dispersion of coordinated fleet in trunk line sections.

[0129] Among the trunk line features mentioned above, features (1) to (5) are macroscopic features, and features (6) to (10) are mesoscopic features.

[0130] Step S24: compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototypes in the trunk line prototype scene library, and determine the scene category of the trunk line to be identified according to the comparison results.

[0131] (1) Red wave scene recognition

[0132] The features required for red wave scene recognition include the above features (1) to (6) and feature (9), a total of 7 features, such as Figure 6 As shown, the specific identification process includes the following steps:

[0133] For each traffic flow feature required for red wave scene recognition, all red wave scene trunk line prototypes / non-red wave scene trunk line prototypes are traversed, and the differences between the trunk line to be recognized and all red wave scene trunk line prototypes / non-red wave scene trunk line prototypes under the same traffic flow feature are calculated respectively;

[0134] After all red wave scene trunk line prototypes / non-red wave scene trunk line prototypes are traversed, the average difference between the trunk line to be identified and all red wave scene trunk line prototypes / non-red wave scene trunk line prototypes under the same traffic flow characteristics is calculated;

[0135] If the average difference between the trunk line to be identified and all trunk line prototypes in red wave scenarios under the same traffic flow feature is less than the average difference between the trunk line to be identified and all trunk line prototypes in non-red wave scenarios under the same traffic flow feature, the traffic flow feature is determined to be a red wave feature; otherwise, the traffic flow feature is determined to be a non-red wave feature;

[0136] After all the traffic flow features required for red wave scene recognition are traversed, the number of red wave features and non-red wave features are counted respectively;

[0137] If the number of red wave features is greater than the number of non-red wave features, the trunk line to be identified is determined to be a red wave scene trunk line; otherwise, the trunk line to be identified is determined to be a non-red wave scene trunk line.

[0138] If the trunk line to be identified is identified as a non-red wave scene trunk line, non-red wave scene identification is further performed.

[0139] (2) Non-red wave scene recognition

[0140] Further scene recognition is performed on the trunk lines with identification results as non-red wave scenarios. Evaluation is performed from multiple dimensions such as the queuing degree of trunk traffic flow, the degree of fleet dispersion, the degree of external interference, etc., and compared with the non-red wave scenario trunk line prototypes of different dimensions through the corresponding features and trunk feature distance measurement method. The more similar scenes are taken as the identification results. Finally, the trunk lines to be identified as non-red wave scenario trunk lines are further identified as states such as continuous queuing / slight queuing, fleet dispersion / tight fleet, and large external interference / small external interference.

[0141] For example: the trunk line prototype scenario library includes 30 non-red wave scenario trunk line prototypes, which are divided into 10 continuous queuing trunk line prototypes and 20 slight queuing trunk line prototypes from the dimension of queuing degree. The traffic flow feature corresponding to the queuing degree is the first feature, and the average difference between the trunk line to be identified and the 10 continuous queuing trunk line prototypes under the first feature, as well as the average difference between the trunk line to be identified and the 20 slight queuing trunk line prototypes under the first feature are obtained respectively; if the average difference between the trunk line to be identified and the 10 continuous queuing trunk line prototypes under the first feature is less than the average difference between the trunk line to be identified and the 20 slight queuing trunk line prototypes under the first feature, then the trunk line to be identified is determined to be in a continuous queuing state; otherwise, the trunk line to be identified is determined to be in a slight queuing state.

[0142] For the two dimensions of discreteness and external interference, they can also be identified in the same way as the above-mentioned queuing degree dimension, and the trunk line to be identified can be identified as a state of discrete fleet or dense fleet, and a state of large external interference or small external interference.

[0143] The three dimensions of queuing degree, discreteness and external interference are explained below.

[0144] ① Queuing degree

[0145] Taking the number of periodic stops of the coordinated fleet on the trunk line section (feature (8)) as the identification feature, based on the trunk feature distance measurement method, the average distance between this feature of the trunk line to be identified and the prototype of the trunk line with slight queues and the prototype of the trunk line with continuous queues is calculated respectively. The trunk line prototype with a closer average distance is taken as the scene recognition result of the non-red wave trunk line to be identified in the queuing degree dimension.

[0146] ② Fleet dispersion

[0147] Taking the degree of discreteness of coordinated convoys on trunk sections (feature (10)) as the identification feature, based on the trunk feature distance measurement method, the average distance between this feature of the trunk to be identified and the prototype of the dense convoy and the prototype of the discrete convoy is calculated respectively. The trunk prototype with a closer average distance is taken as the scene recognition result of the non-red wave trunk to be identified in the convoy discreteness dimension.

[0148] ③External interference degree

[0149] Taking the non-coordinated flow rate in the upstream direction of the trunk line section (feature (7)) as the identification feature, based on the trunk feature distance measurement method, the average distance between this feature of the trunk line to be identified and the prototype of the trunk line with large external interference and the prototype of the trunk line with small external interference are calculated respectively, and the prototype scene with a closer average distance is taken as the scene recognition result of the non-red wave trunk line to be identified in the external interference dimension.

[0150] In step S24, the trunk feature difference measurement is the key core part. Among the 10 trunk features extracted in step S23, features (1) to (4) and feature (9) are numerical features, feature (5) is a distribution type, and the remaining features are listed in units of trunk segments. Different types of features have different difference calculation methods:

[0151] ① Numerical features

[0152] The difference between single-valued features between trunks can be measured by the absolute value of the difference, calculated as follows:

[0153]

[0154] in, The single value feature f is the relationship between trunk a and trunk b. value The difference in is the eigenvalue of trunk a, is the eigenvalue of trunk line b.

[0155] ②Distribution characteristics

[0156] The distribution of travel time and travel speed on trunk lines belongs to the distribution characteristics, which is measured by the Wasserstein / Earth-Mover Distance. The expression method is as follows:

[0157]

[0158] in, is the distribution characteristics f of trunk a and trunk b pd The difference in is the characteristic distribution of trunk a, is the characteristic distribution of trunk b, and w(·) is the distribution distance calculation function.

[0159] ③Road section list type features

[0160] For the trunk line segment list type feature, the difference between two trunk lines is measured by considering the different number of trunk line segments and the upstream and downstream connection relationship and continuity of the trunk line itself. Suppose there is a long trunk line a with m segments and a short trunk line b with n segments (m≥n). The feature difference method is as follows: Figure 7The specific measurement method is as follows:

[0161] Starting from the starting points of trunk a and trunk b, the differences of the road segment groups with a length of n are compared as a whole, and a total of m-n+1 comparisons are made;

[0162] When comparing the differences between the road section groups of trunk a and trunk b, the characteristic differences of a pair of road sections are calculated one by one in the order from upstream to downstream of the two trunk sections, and the calculation is performed n times in total. Finally, the average difference of the n road section pairs is taken as the difference of the road section group.

[0163] The average difference of m-n+1 road segment groups is calculated as the difference in the characteristics of the two trunk lines.

[0164] The method for measuring the feature differences between road sections is determined by the element structure in the trunk list: the road section features in features (6) to (7) are time series, and the dynamic time warping (DTW) algorithm is used to measure the feature differences between road sections; the road section features in features (8) and (10) are numerical values, and the absolute value of the difference between the two numerical values ​​is used to calculate the road section feature differences.

[0165] Step S25: Select a corresponding coordination control strategy according to the determined scene category of the trunk line to be identified.

[0166] According to the trunk line scene recognition results shown in Table 1, the corresponding control strategy is selected from the trunk line coordination control strategy recommendation table shown in Table 2 to achieve the purpose of optimizing the trunk line control. Specifically, it may include:

[0167] If the scene recognition result is a red wave scene, the red wave coordinated control strategy is selected;

[0168] If the scene recognition result is a non-red wave scene and there is a continuous queue, the congestion coordination control strategy is selected;

[0169] If the scene recognition result is a non-red wave scene, and there is a slight queue and large external interference, the green light terminal type green wave coordinated control strategy is selected;

[0170] If the scene recognition result is a non-red wave scene, and there is a slight queue, little external interference, and a tight fleet, the green light starting point type green wave coordinated control strategy is selected;

[0171] If the scene recognition result is a non-red wave scene, and there is a slight queue, little external interference, and a discrete fleet, the green light center type green wave coordinated control strategy is selected.

[0172] Table 1 Trunk scene recognition results

[0173]

[0174] Table 2 Recommended trunk line coordination control strategies

[0175] Scene recognition results Recommended control strategy Red wave scene Red Wave Coordination Non-red wave scenario: continuous queueing Congestion coordination Non-red wave scenario: slight queue, large external interference Green light terminal type green wave coordination Non-red wave scenarios: slight queues, little external interference, and tight convoys Green light starting point green wave coordination Non-red wave scenarios: slight queues, little external interference, and scattered fleets Green light center type green wave coordination

[0176] The trunk line scene recognition method provided by the embodiment of the present invention obtains multiple traffic flow features associated with the trunk line to be identified and the trunk line coordinated control scene based on the static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated section, and identifies whether the trunk line belongs to the red wave control scene by comparing the feature differences between the trunk line to be identified and multiple different trunk line prototypes in the trunk line prototype scene library, and can evaluate the traffic status of the non-red wave scene trunk line from multiple dimensions such as queuing degree, fleet dispersion degree, external interference degree, etc. Therefore, the scene category of the trunk line to be identified can be determined more finely, providing a good decision-making basis for trunk line coordinated control.

[0177] The embodiment of the present invention also provides a trunk scene recognition device, such as Figure 8 As shown, the device comprises:

[0178] An acquisition unit 11 is used to acquire static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated road sections during the target period;

[0179] An extraction unit 12 is used to extract multiple traffic flow features of the trunk to be identified according to the static road network information and dynamic traffic flow information of the trunk to be identified and the trunk-associated road sections during the target period;

[0180] The identification unit 13 is used to compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determine the scene category of the trunk line to be identified according to the comparison result.

[0181] The trunk line scene recognition device provided in the embodiment of the present invention obtains multiple traffic flow characteristics associated with the trunk line to be identified and the trunk line coordination control scene based on the static road network information of the trunk line to be identified and the trunk line associated road sections and the dynamic traffic flow information, and determines the scene category of the trunk line to be identified by comparing the feature differences between the trunk line to be identified and multiple different trunk line prototypes in the trunk line prototype scene library, thereby being able to more finely determine the scene category of the trunk line to be identified, providing a good decision-making basis for trunk line coordination control.

[0182] Optionally, the device further comprises:

[0183] A construction unit is used to construct a trunk prototype scenario library before the acquisition unit 11 acquires the static road network information and dynamic traffic flow information of the trunk to be identified and the trunk-associated sections during the target period, wherein the trunk prototype scenario library includes multiple red wave scenario trunk prototypes and multiple non-red wave scenario trunk prototypes, and each red wave scenario trunk prototype and each non-red wave scenario trunk prototype correspond to multiple traffic flow characteristics respectively.

[0184] Optionally, the identification unit 13 is also used to compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of all red wave scene trunk line prototypes and all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and determine whether the trunk line to be identified is a red wave scene trunk line based on the comparison result.

[0185] Optionally, the identification unit 13 includes:

[0186] The first acquisition module is used to obtain, for each traffic flow feature required for red wave scene recognition, the average difference between the trunk line to be identified and all red wave scene trunk line prototypes under the same traffic flow feature, and the average difference between the trunk line to be identified and all non-red wave scene trunk line prototypes under the same traffic flow feature;

[0187] A determination module, for determining that the traffic flow feature is a red wave feature when the average difference between the trunk line to be identified and all trunk line prototypes of red wave scenarios under the same traffic flow feature is less than the average difference between the trunk line to be identified and all trunk line prototypes of non-red wave scenarios under the same traffic flow feature; otherwise, determining that the traffic flow feature is a non-red wave feature;

[0188] A statistical module is used to count the number of red wave features and non-red wave features respectively;

[0189] The first identification module is used to determine that the trunk line to be identified is a red wave scene trunk line when the number of red wave features is greater than the number of non-red wave features; otherwise, determine that the trunk line to be identified is a non-red wave scene trunk line.

[0190] Optionally, the non-red wave scenario trunk line prototype is divided into multiple dimensions, wherein the multiple dimensions include at least two of the following: queuing degree, fleet dispersion degree, and external interference degree, and each dimension is divided into two different non-red wave scenario trunk line prototypes;

[0191] The identification unit is also used to compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of all non-red wave scenario trunk line prototypes in the trunk line prototype scene library for each dimension when it is determined that the trunk line to be identified is a non-red wave scenario trunk line, and determine the non-red wave scenario category of the trunk line to be identified in the dimension according to the comparison result.

[0192] Optionally, the identification unit 13 further includes:

[0193] The second acquisition module is used to respectively obtain the average difference between the trunk line to be identified and all the first non-red wave scenario trunk line prototypes under the dimension under the traffic flow characteristics corresponding to the dimension, and the average difference between the trunk line to be identified and all the second non-red wave scenario trunk line prototypes under the dimension under the traffic flow characteristics corresponding to the dimension;

[0194] The second identification module is used to determine that the trunk line to be identified is the first non-red wave scene trunk line under the dimension when the average difference between the trunk line to be identified and all the first non-red wave scene trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension is less than the average difference between the trunk line to be identified and all the second non-red wave scene trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension; otherwise, determine that the trunk line to be identified is the second non-red wave scene trunk line under the dimension.

[0195] Optionally, the device further comprises:

[0196] A selection unit is used for comparing the traffic flow characteristics of the trunk line to be identified in the identification unit 13 with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and after determining the scene category of the trunk line to be identified according to the comparison result, selecting the corresponding coordination control strategy according to the determined scene category of the trunk line to be identified; if the scene category of the trunk line to be identified is a red wave scene, the red wave coordination control strategy is selected; if the scene category of the trunk line to be identified is a non-red wave scene, the congestion coordination control strategy or the green wave coordination control strategy is selected according to the non-red wave scene category of the trunk line to be identified.

[0197] Specifically, if the scene category of the trunk line to be identified is a red wave scene, the red wave coordination control strategy is selected; if the scene category of the trunk line to be identified is a non-red wave scene, the congestion coordination control strategy or the green wave coordination control strategy is selected according to the non-red wave scene category of the trunk line to be identified in each dimension, wherein the non-red wave scene trunk prototype is divided into multiple dimensions.

[0198] The device of this embodiment can be used to execute the technical solution of the above method embodiment. Its implementation principle and technical effect are similar and will not be repeated here.

[0199] An embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0200] at least one processor; and

[0201] a memory communicatively connected to the at least one processor; wherein,

[0202] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above-mentioned trunk scene recognition method.

[0203] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the above-mentioned trunk scene recognition method is implemented.

[0204] Those skilled in the art can understand that all or part of the processes in the above method embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes in the above method embodiments. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0205] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A trunk scene recognition method, characterized in that: The method comprises: Obtain static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line-related sections during the target period; Extracting multiple traffic flow features of the trunk line to be identified based on the static road network information and dynamic traffic flow information of the trunk line to be identified and the road sections associated with the trunk line during the target period; The traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of the trunk line prototypes in the trunk line prototype scene library, and the scene category of the trunk line to be identified is determined according to the comparison result.

2. The method according to claim 1, characterized in that Before acquiring the static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated road sections during the target period, the method further includes: A trunk line prototype scenario library is constructed, wherein the trunk line prototype scenario library includes a plurality of red wave scenario trunk line prototypes and a plurality of non-red wave scenario trunk line prototypes, and each red wave scenario trunk line prototype and each non-red wave scenario trunk line prototype correspond to a plurality of traffic flow features respectively.

3. The method according to claim 1, characterized in that Comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determining the scene category of the trunk line to be identified according to the comparison result, including: The traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of all red wave scenario trunk line prototypes and all non-red wave scenario trunk line prototypes in the trunk line prototype scenario library, and whether the trunk line to be identified is a red wave scenario trunk line is determined according to the comparison results.

4. The method according to claim 3, characterized in that The step of comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of all red wave scene trunk line prototypes and all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and determining whether the trunk line to be identified is a red wave scene trunk line according to the comparison result comprises: For each traffic flow feature required for red wave scene recognition, obtain the average difference between the trunk line to be recognized and all trunk line prototypes of red wave scenes under the same traffic flow feature, and the average difference between the trunk line to be recognized and all trunk line prototypes of non-red wave scenes under the same traffic flow feature; If the average difference between the trunk line to be identified and all trunk line prototypes in red wave scenarios under the same traffic flow feature is less than the average difference between the trunk line to be identified and all trunk line prototypes in non-red wave scenarios under the same traffic flow feature, the traffic flow feature is determined to be a red wave feature; otherwise, the traffic flow feature is determined to be a non-red wave feature; Count the number of red wave features and non-red wave features respectively; If the number of red wave features is greater than the number of non-red wave features, the trunk line to be identified is determined to be a red wave scene trunk line; otherwise, the trunk line to be identified is determined to be a non-red wave scene trunk line.

5. The method according to claim 3 or 4, characterized in that: The non-red wave scenario trunk line prototype is divided into multiple dimensions, wherein the multiple dimensions include at least two of: queuing degree, fleet dispersion degree and external interference degree, and each dimension is divided into two different non-red wave scenario trunk line prototypes; If it is determined that the trunk line to be identified is a non-red wave scene trunk line, then comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determining the scene category of the trunk line to be identified according to the comparison result also includes: For each dimension, the traffic flow characteristics of the trunk line to be identified are compared with the corresponding traffic flow characteristics of all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and the non-red wave scene category of the trunk line to be identified under the dimension is determined according to the comparison result.

6. The method according to claim 5, characterized in that The step of comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of all non-red wave scene trunk line prototypes in the trunk line prototype scene library, and determining the non-red wave scene category of the trunk line to be identified under the dimension according to the comparison result includes: Respectively obtain the average difference between the trunk line to be identified and all the first non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension, and the average difference between the trunk line to be identified and all the second non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension; If the average difference between the trunk line to be identified and all the first non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension is less than the average difference between the trunk line to be identified and all the second non-red wave scenario trunk line prototypes under the dimension in the traffic flow characteristics corresponding to the dimension, then the trunk line to be identified is determined to be the first non-red wave scenario trunk line under the dimension; otherwise, the trunk line to be identified is determined to be the second non-red wave scenario trunk line under the dimension.

7. The method according to any one of claims 1 to 6, characterized in that The static road network information and dynamic traffic flow information of the trunk road section to be identified and the trunk-associated road section during the target period include at least one of the following information: section length, direction, free flow speed, upstream and downstream topological relationship, flow in each direction, saturation in each direction, green light time of each lane in the section and vehicle passing record.

8. The method according to any one of claims 1 to 6, characterized in that After comparing the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library and determining the scene category of the trunk line to be identified according to the comparison result, the method further includes: According to the determined scene category of the trunk line to be identified, select the corresponding coordination control strategy, wherein if the scene category of the trunk line to be identified is a red wave scene, select the red wave coordination control strategy; if the scene category of the trunk line to be identified is a non-red wave scene, select the congestion coordination control strategy or the green wave coordination control strategy according to the non-red wave scene category of the trunk line to be identified.

9. A trunk scene recognition device, characterized in that: The device comprises: An acquisition unit, used to acquire static road network information and dynamic traffic flow information of the trunk line to be identified and the trunk line associated road sections during the target period; An extraction unit, configured to extract a plurality of traffic flow features of the trunk line to be identified according to the static road network information and dynamic traffic flow information of the trunk line to be identified and the road sections associated with the trunk line during the target period; The identification unit is used to compare the traffic flow characteristics of the trunk line to be identified with the corresponding traffic flow characteristics of the trunk line prototype in the trunk line prototype scene library, and determine the scene category of the trunk line to be identified according to the comparison result.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images