A method for judging traffic status at urban arterial intersections

By preprocessing and real-time analysis of vehicle data at the intersection of urban trunk lines, we can determine whether the intersection is in an oversaturated, overloaded or imbalanced state, solving the problem of low utilization rate of vehicle identification data, and improving the refined level and intersection of traffic management Port traffic efficiency.

CN118629231BActive Publication Date: 2024-11-29CHINA RAILWAY ERYUAN ENGINEERING GROUP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202410735752.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-07
Publication Date
2024-11-29
Estimated Expiration
2044-06-07

AI Technical Summary

Technical Problem

In the prior art, the utilization rate of vehicle identification data is low, and it is impossible to effectively assist in real-time analysis and management of traffic status at the intersection of urban trunk lines, resulting in difficult to solve the problem of traffic congestion.

Method used

By collecting and preprocessing the historical and real-time vehicle data of the target intersection, calculate the historical judgment threshold of the supersaturation state, determine whether it belongs to the supersaturation cycle, and determine whether it is overloaded or unbalanced during the supersaturation cycle, and output the judgment result to achieve the intersection Accurate identification and management of traffic conditions.

Benefits of technology

It realizes the effective utilization of massive vehicle identification data, provides a basis for real-time traffic state analysis and signal timing optimization, and improves the traffic efficiency at the intersection and the refined management level of road traffic.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118629231B_ABST
    Figure CN118629231B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for distinguishing the traffic status of an urban trunk intersection, comprising the following steps: collecting historical vehicle data of a target intersection, preprocessing the data, and calculating a historical threshold for distinguishing an oversaturated state; collecting real-time vehicle data of each cycle of the target intersection, preprocessing the data, and judging whether it belongs to an oversaturated cycle based on the historical threshold for distinguishing an oversaturated state; if it belongs to an oversaturated cycle, judging whether the target intersection is overloaded within the oversaturated cycle; if the target intersection is not overloaded, judging whether the target intersection is unbalanced within the oversaturated cycle; and outputting the judgment result. The present invention provides a method for distinguishing the traffic status of an urban trunk intersection to solve the problem of low utilization rate of a large amount of vehicle identification data in the prior art, and to achieve the purpose of fully tapping the stock data resources, conducting real-time analysis of the traffic status of the intersection through vehicle identification data, and improving the level of refined management of road traffic.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of urban traffic management, and in particular to a method for determining the traffic status of an urban trunk line intersection. Background Art

[0002] With the rapid growth of motor vehicle traffic demand, traffic congestion in large cities is getting worse; the state identification, cause analysis and optimization control of oversaturated signal-controlled intersections have become important directions for theoretical research and practice in the field of traffic signal control. Accurate, comprehensive and real-time traffic flow detection data is the basis for theoretical research on traffic flow and optimization of intersection signal control. In recent years, with the continuous improvement of image recognition, sensor control, information transmission and storage technologies, automatic license plate recognition systems have been widely used. Automatic license plate recognition systems can not only collect information such as license plate numbers and passing times from passing vehicle image information, but also obtain information such as vehicle paths and road section travel time by matching license plate data at different detection points. Compared with other detection methods, the obtained license plate recognition data has the advantages of strong continuity, high recognition rate, large sample size, wide coverage, high stability and strong time continuity. With the continuous improvement and deepening of the construction of intelligent transportation systems in large cities, the coverage rate of automatic license plate recognition systems has continued to increase. In some cities, coverage of all signal-controlled intersections on main roads has been achieved, and the basic conditions for using license plate data to extract traffic status parameters and optimize signal timing plans have been met.

[0003] Based on the license plate recognition system, traffic management departments and safety management departments will obtain and store a large amount of data on vehicles passing through urban arterial intersections. However, the utilization rate of such data in existing technologies is low. It is generally only used for the identification of vehicle violations and public safety law enforcement needs, and cannot contribute to the real-time analysis and management of intersection traffic conditions. Therefore, it is of great theoretical significance and technical application value to explore how to use the current massive license plate data to describe road traffic conditions, analyze the causes of intersection oversaturation, and optimize signal timing schemes. Summary of the invention

[0004] The present invention provides a method for determining the traffic status of an urban trunk intersection, so as to solve the problem of low utilization rate of a large amount of vehicle identification data in the prior art, realize the purpose of fully tapping into the stock data resources, performing real-time analysis of the traffic status of the intersection through vehicle identification data, and improving the level of refined management of road traffic.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for determining the traffic status of an urban trunk road intersection comprises the following steps:

[0007] Collect historical vehicle data of the target intersection, preprocess the data, and calculate the historical threshold for oversaturation status;

[0008] Collect real-time vehicle data of each cycle at the target intersection, pre-process the data, and determine whether it belongs to an oversaturation cycle based on the oversaturation state historical discrimination threshold;

[0009] If it is an oversaturation period, then during the oversaturation period, determine whether the target intersection is overloaded;

[0010] If the target intersection is not overloaded, then during the oversaturation period, it is determined whether the target intersection is unbalanced;

[0011] Output the judgment result.

[0012] In view of the problem in the prior art that the utilization rate of vehicle identification data is low and it cannot assist in the identification and management of the traffic status at intersections, the present invention proposes a method for determining the traffic status at urban trunk intersections.

[0013] The intersection is the gathering place of traffic flow in all directions. Affected by signal control, the continuous flow of vehicles is interrupted. It is one of the core points that cause traffic congestion on urban trunk roads. Real-time judgment of the traffic status of the intersection is conducive to improving the management level of the intersection in the background, thereby improving the traffic efficiency of the intersection and improving the level of refined management of road traffic. Specifically, this application first pre-processes the historical vehicle data of the target intersection, and then calculates the historical judgment threshold of the oversaturation state for standby; then collects the real-time vehicle data of each cycle of the target intersection, and after each cycle, pre-processes the data and determines whether it belongs to the oversaturation cycle; if it belongs to the oversaturation cycle, continue to judge whether the target intersection is overloaded during the oversaturation cycle; if so, output the judgment result; if not, continue to judge whether the target intersection is unbalanced, and output the final judgment result.

[0014] Those skilled in the art should understand that the oversaturation of an intersection refers to the state where the traffic volume arriving at an intersection exceeds its capacity. For an unsaturated cycle, the number of vehicles arriving in the cycle is equal to the number of vehicles leaving the cycle, that is, the volume passing the stop line is equal to the traffic demand; while for an oversaturated cycle, the volume passing the stop line only represents the capacity of the intersection, which is not equal to the number of vehicles arriving in the cycle, nor can it represent the traffic demand borne by the entrance road in the cycle; in addition, the cycle in this application refers to the control cycle of the intersection signal light.

[0015] It can be seen that the present application can make full use of the existing license plate recognition system of the traffic management department or the safety management department, realize the effective use of the massive stock data in the prior art, and can perform real-time analysis and identification of each cycle of the road intersection, identify the traffic operation status of the current cycle, and provide a more sufficient and scientific basis for the dynamic control of the road intersection and the dynamic optimization of the signal timing scheme, which is conducive to improving the traffic efficiency of the intersection and improving the level of refined management of road traffic.

[0016] In addition, the present application can not only identify oversaturation periods, but also determine whether the target intersection is overloaded or unbalanced during the oversaturation period, thereby achieving more precise and accurate traffic status identification.

[0017] Furthermore, the data preprocessing includes data reduction, data conversion, and data cleaning performed in sequence;

[0018] The data reduction method includes: retaining the license plate data, the vehicle entrance lane data, and the vehicle lane data in the collected vehicle data, and deleting the remaining data;

[0019] The data conversion method includes: assigning lane directions to all lanes of each entrance road of the target intersection, and superimposing the lane directions onto the collected vehicle data;

[0020] The data cleaning includes executing in sequence: cleaning duplicate data and eliminating abnormal data.

[0021] In the existing technology, the technology for obtaining vehicle data mainly includes vehicle identification technology used by traffic or safety management departments. The core of this type of existing technology is license plate recognition. The original license plate data obtained is prone to problems such as data duplication, data anomalies, and data missing. Therefore, it is necessary to pre-process the original data before conducting subsequent data analysis and mining.

[0022] The data preprocessing process in this solution is applicable to both the preprocessing of historical vehicle data and the preprocessing of real-time vehicle data in each cycle.

[0023] First of all, the vehicle data collected by the prior art includes a large number of fields for each vehicle's passing data, and only some of the fields contribute to the discrimination method of the present application. Therefore, in the preprocessing process, the present scheme first performs data reduction, retains the license plate data, the vehicle's entrance lane data, and the vehicle's lane data in a large number of fields, and deletes the remaining data, thereby reducing the amount of data and ensuring that the loss of information is minimized.

[0024] Then the data is converted. In the license plate recognition data, the intersection number, vehicle driving direction and lane number represent the spatial attribute information of the vehicle. For the target intersection, the intersection number is fixed and known; the vehicle driving direction can only reflect the direction of the vehicle at the time of detection, that is, the direction of the entrance road of the intersection in which the vehicle is located; the lane number can only reflect which lane the vehicle is located at the time of detection. Due to the lack of the lane function of the lane where the vehicle is located, the direction of the detected vehicle after passing through the intersection cannot be directly inferred using this type of known data. Therefore, this scheme assigns lane directions to all lanes of each entrance road of the target intersection through data conversion, and superimposes the lane direction on the collected vehicle data, so that when the vehicle is successfully detected by the detection equipment, the direction of the entrance road and the lane function of the lane where the vehicle is located are defined, and then the number of the next intersection that the detected vehicle passes through after passing the detection intersection can be inferred.

[0025] After data conversion, this solution also performs data cleaning on the collected vehicle data. The data cleaning process includes cleaning duplicate data and eliminating abnormal data in sequence.

[0026] Furthermore, the method for cleaning up duplicate data includes:

[0027] Extract all license plate data of an entrance lane of the target intersection, group the extracted license plate data by license plate type, and sort the license plate data in each group by collection time;

[0028] Divide each group of sorted license plate data into a number of first time intervals according to the first time interval, and determine whether there are duplicate license plates in each first time interval: if so, retain the last vehicle data of the duplicate license plate sorted by time, and delete the remaining vehicle data of the duplicate license plate;

[0029] Extract all license plate data of the next entrance road of the target intersection until all entrance roads of the target intersection have been cleared of duplicate data;

[0030] The method for removing abnormal data includes: deleting license plate data with missing license plate recognition and deleting license plate data with license plate character recognition errors.

[0031] Duplicate data in license plate recognition data refers to data in which the same vehicle is detected multiple times at the same location within a small time interval; among them, duplicate data generated by the same vehicle being repeatedly captured multiple times by the same device within a small time interval is relatively easy to clean; however, there are also some duplicate data, which are generated by the same vehicle being repeatedly captured multiple times by different devices on the same entrance road within a small time interval. The cleaning of such duplicate data is relatively difficult. According to the experiments of the inventor's team, such duplicate data accounts for about 0.4% to 0.5%, and it is still necessary to clean it. In order to overcome the above problems, this scheme first extracts all license plate data of an entrance road of the target intersection, groups the extracted license plate data by license plate type, and sorts them by collection time in each group of license plate data, and then divides them into several first time intervals according to the first time interval, and determines whether there are duplicate license plates in each first time interval, thereby deleting duplicate data, and performing the above operations on all entrance roads of the target intersection until all entrance roads have completed the steps of cleaning duplicate data of this sub-scheme. It can be seen that this solution cleans up duplicate data based on the inlet channel, so it can effectively overcome the duplicate data generated by repeated capture by different devices on the same inlet channel and improve the effect of data cleaning.

[0032] In addition, due to ambient lighting, bad weather, damaged or faded license plates, equipment failure and other reasons, license plate recognition data may be abnormal, manifested as missing characters, garbled errors or complete recognition failure. This solution deletes license plate data with missing license plate recognition, that is, first eliminates license plate data with clear recognition failures; then deletes license plate data with license plate character recognition errors, that is, continues to eliminate license plate data with non-compliant license plate characters, to further ensure the data accuracy of subsequent operations.

[0033] Furthermore, the method for calculating the oversaturation state history discrimination threshold includes:

[0034] dividing the preprocessed historical vehicle data into a plurality of second time intervals according to a second time interval;

[0035] In the second most recent time interval, the hourly flow rate and equivalent density of the target intersection are calculated, and an hourly flow rate-equivalent density map is established;

[0036] Based on the hourly flow rate-equivalent density diagram, determine whether there is an oversaturation state in the second time interval:

[0037] If so, obtaining a supersaturation state historical discrimination threshold based on the hourly flow rate-equivalent density diagram in the second time interval;

[0038] If not, the most recent second time interval is replaced by the previous second time interval, and the process is iterated forward until the oversaturation state historical judgment threshold is obtained.

[0039] This scheme proposes a method to determine the historical discrimination threshold of the oversaturation state based on the idea of ​​equivalent density. This method can make full use of the stock data to determine the historical discrimination threshold of the oversaturation state in the current real-time discrimination process, and dynamically update the historical discrimination threshold of the oversaturation state according to different discrimination times.

[0040] The most recent second time interval mentioned in this solution refers to a second time interval closest to the currently determined period; the forward iteration mentioned in this solution refers to replacing the current second time interval with the previous second time interval in the time dimension.

[0041] Furthermore, the equivalent density is calculated by the following formula: Where: D is equivalent density; hV is hourly flow rate; L is average section length; T is average travel time of section;

[0042] The method for establishing an hourly flow rate-equivalent density diagram comprises:

[0043] A data set is established with the obtained hourly flow rate and equivalent density, wherein the data set includes a plurality of scattered points; wherein each scattered point includes its corresponding {equivalent density, hourly flow rate} data;

[0044] For all scattered points, a sliding average is performed with a specified sliding window and a specified step size to obtain a number of sliding scattered points; the sliding scattered points include {smoothed density, smoothed hourly flow rate} data;

[0045] A two-dimensional coordinate system is established with equivalent density as the horizontal coordinate and hourly flow rate as the vertical coordinate; all the scattered points after sliding are marked in the two-dimensional coordinate system;

[0046] The method for determining whether an oversaturation state exists within the second time interval includes:

[0047] Perform linear fitting of a single segment straight line on all the scattered points after sliding in the two-dimensional coordinate system, and calculate the first root mean square error based on the fitting result;

[0048] Perform piecewise linear fitting of two straight lines on all the scattered points after sliding in the two-dimensional coordinate system, and calculate the second root mean square error based on the fitting results;

[0049] Compare the first RMS error to the second RMS error:

[0050] If the first root mean square error is greater than the second root mean square error, an oversaturation state exists in the corresponding second time interval;

[0051] If the first root mean square error is less than or equal to the second root mean square error, then there is no oversaturation state in the corresponding second time interval;

[0052] The method for obtaining the oversaturation state historical discrimination threshold is: defining the smooth density corresponding to the intersection point of two straight lines in the result of the piecewise linear fitting as the critical density; and calculating the oversaturation state historical discrimination threshold = 1.2×critical density.

[0053] In traditional technology, the judgment of oversaturation is generally based on the demand flow ratio. However, since the actual demand needs to be estimated through the number of vehicle queues, there is an inevitable lag in the analysis of oversaturation. Therefore, this method is only applicable to the retrospective review of the congestion formation and dissipation process, and is not applicable to the real-time judgment of the traffic status of the intersection. The method for judging the oversaturation state proposed in this scheme can meet the real-time demand, and the inventor has experimentally verified that the judgment result of the oversaturation period is basically consistent with the traditional viewpoint demand flow ratio method (only 1 to 2 minutes of error). Therefore, the historical judgment threshold of the oversaturation state obtained based on this method can fully meet the needs of this application.

[0054] Furthermore, the method for judging whether it is an oversaturation period based on the oversaturation state history judgment threshold includes:

[0055] Based on the pre-processed real-time vehicle data of the target intersection in each cycle, the equivalent density in the cycle is calculated in real time; the calculation method of the equivalent density is the same as the equivalent density in the process of calculating the historical discrimination threshold of the oversaturation state, which will not be described in detail here;

[0056] Compare the equivalent density within this period with the historical threshold for supersaturation:

[0057] If the equivalent density in the cycle is greater than or equal to the historical discrimination threshold of the oversaturation state, the cycle belongs to the oversaturation cycle;

[0058] If the equivalent density in this period is less than the historical judgment threshold of the oversaturation state, then this period does not belong to the oversaturation period.

[0059] Further, the method for determining whether the target intersection is overloaded includes:

[0060] Calculate the saturation flow rate and cycle flow of each cycle at the target intersection;

[0061] Calculate the flow ratio of each entrance of the target intersection separately: In the formula, y i is the flow ratio of the i-th entrance of the target intersection; q i is the cycle flow of the i-th entrance of the target intersection; s i is the saturated flow rate of the i-th entrance of the target intersection;

[0062] Calculate the sum of the flow ratios of all entrances at the target intersection: Y = ∑y i ;

[0063] Calculate the overload factor η: Where: λ is the green-to-signal ratio of the cycle; G is the total effective green light time in the current cycle; C is the cycle length;

[0064] Determine η: If η>1, the target intersection is overloaded.

[0065] Intersection overload means that over a period of time, the traffic volume borne by the intersection has exceeded its capacity. It is the direct result of traffic demand far exceeding traffic supply. When the traffic volume is large, the traffic-congested intersection will turn into an overloaded state, and even cause the intersection to be locked.

[0066] In this solution, the saturation flow rate is estimated in advance; the cycle flow is calculated in real time, which means that after each cycle, the real-time vehicle data of each cycle is collected and calculated. This solution clearly defines the specific method of judging whether the target intersection is overloaded based on the overload coefficient, which is more conducive to the real-time judgment of the traffic status of the intersection.

[0067] Furthermore, in the process of calculating the periodic flow of each cycle at the target intersection, by extracting the vehicle model in the vehicle data, the standard flow of different vehicle models is converted according to the following formula: Pcu=α×Veh; where Pcu is the standard vehicle flow; Veh is the unit vehicle flow; α is the vehicle model coefficient.

[0068] In the extracted vehicle data, the traffic flow Veh is the natural flow. However, since the vehicle model, length and other factors directly affect its occupation of road resources, this scheme converts the traffic flow into standard traffic flow through the vehicle model coefficient α to improve the accuracy of judging whether the target intersection is overloaded.

[0069] The saturation flow rate of the i-th entrance of the target intersection is calculated by the following method:

[0070] Based on the historical vehicle data of the target intersection, identify at least N saturated fleets with different periods, where N ≥ 50;

[0071] Get a set of headway time sequences in each saturated fleet;

[0072] Eliminate the first four headway times in each group of headway time sequences to obtain the mean headway time of the remaining groups;

[0073] The average of all the single-group remaining headway means is calculated to obtain the average saturated headway, and the saturated flow rate is calculated based on the average saturated headway.

[0074] Among them, the flow rate calculation based on the headway time is an existing commonly used algorithm, which will not be elaborated here.

[0075] The calculation of saturated flow in the prior art mainly considers the correction coefficient on the basis of the ideal saturated flow rate of the lane. The problem with this calculation method is that the acquisition accuracy of various correction coefficients is low, resulting in low accuracy of the calculation results. Based on this, this scheme proposes a method for calculating the saturated flow rate based on real-time collected vehicle data. This method can perform separate calculations for each entrance lane, and by extracting the idea of ​​the saturated fleet and determining the head vehicle of the saturated fleet, the calculation accuracy of the saturated flow rate is significantly improved. For different intersections, and even for different phases and different lanes of the same intersection, the front loss time is different, so this scheme eliminates the first 4 headway time sequences in each group of headway time sequences to reduce the temporal and spatial differences in the calculation results.

[0076] Furthermore, the saturated fleet is identified by the following method:

[0077] Identify the leading vehicle in a convoy: extract the vehicle whose headway is greater than t+20s and whose headway of the eight vehicles that pass in succession is less than or equal to 6s as the leading vehicle in the convoy; where t is the red light duration of the corresponding lane of the entrance lane;

[0078] Starting from the head vehicle of the convoy, all vehicles whose headway time is continuously less than or equal to 6s are regarded as a saturated convoy.

[0079] This solution sets the minimum length of a saturated convoy to 8 vehicles and the headway threshold to 6s. This is to take the saturated headway of large vehicles into account and to reflect the characteristics of queued vehicles passing the stop line continuously.

[0080] Further, the method for determining whether the target intersection is unbalanced includes:

[0081] Calculate the flow ratio of each lane in each entrance lane of the target intersection: In the formula, y im is the flow ratio of the mth lane in the i-th entrance of the target intersection; q im is the cycle flow of the mth lane in the i-th entrance of the target intersection; s im is the saturation flow rate of the mth lane in the i-th entrance of the target intersection;

[0082] Calculate the green signal ratio λ of the mth lane in the i-th entrance lane im : In the formula, g im is the effective green light time corresponding to the mth lane in the i-th entrance lane; C is the cycle length;

[0083] Judgment: If any y im >λ im , then the target intersection is unbalanced.

[0084] Intersection imbalance refers to the phenomenon that the delays and saturation of each flow direction at the intersection are seriously differentiated. This application makes an imbalance judgment on the premise that the target intersection is not overloaded, and clearly defines the detailed method steps for judging whether there is an imbalance in this solution, providing data support for further signal optimization.

[0085] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0086] 1. A method for distinguishing the traffic status of an urban trunk intersection of the present invention can make full use of the existing license plate recognition system of the traffic management department or the safety management department, realize the effective use of the massive stock data in the prior art, and can perform real-time analysis and recognition of each cycle of the road intersection, recognize the traffic operation status of the current cycle, and provide a more sufficient and scientific basis for the dynamic control of the road intersection, the dynamic optimization of the signal timing scheme, etc., so as to improve the traffic efficiency of the intersection and improve the refined management level of road traffic.

[0087] 2. The method for distinguishing the traffic status of an urban trunk intersection of the present invention can not only identify the oversaturation period, but also determine whether the target intersection is overloaded or unbalanced within the oversaturation period, thereby achieving more precise and accurate traffic status recognition.

[0088] 3. The present invention provides a method for determining the traffic status of an urban trunk intersection, which cleans up duplicate data based on the entrance lane, thereby effectively overcoming the duplicate data generated by repeated capture by different devices on the same entrance lane, and improving the effect of data cleaning.

[0089] 4. The present invention provides a method for determining the traffic status of an urban trunk intersection, which can make full use of the existing data to determine the historical determination threshold of the oversaturation state in the current real-time determination process, and dynamically update the historical determination threshold of the oversaturation state according to different determination times; the proposed method for determining the oversaturation state can meet real-time needs.

[0090] 5. The present invention provides a method for determining the traffic status of an urban trunk intersection, which can calculate the saturated flow rate for each entrance road separately, and by extracting the idea of ​​a saturated fleet and determining the head vehicle of the saturated fleet, the calculation accuracy of the saturated flow rate is significantly improved.

[0091] 6. The present invention provides a method for determining the traffic status of a city trunk intersection and proposes a method for identifying a saturated vehicle fleet, which is conducive to quickly and accurately calculating the saturated flow rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0093] Figure 1 It is a schematic diagram of a flow chart of a specific embodiment of the present invention;

[0094] Figure 2 It is a linear fitting schematic diagram of a single straight line in a specific embodiment of the present invention;

[0095] Figure 3 It is a schematic diagram of piecewise linear fitting in a specific embodiment of the present invention;

[0096] Figure 4 It is a schematic diagram of the relationship between the periodic flow rate and the average travel time in a specific embodiment of the present invention;

[0097] Figure 5 It is a schematic diagram of the identification process of a saturated vehicle fleet in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0098] In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the examples and the accompanying drawings. The schematic embodiments of the present invention and the description thereof are only used to explain the present invention and are not intended to limit the present invention. In the description of the present application, it should be understood that the orientation or positional relationship indicated by terms such as "front", "rear", "left", "right", "up", "down", "vertical", "horizontal", "high", "low", "inside", "outside", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present application.

[0099] Embodiment 1:

[0100] like Figure 1 A method for determining the traffic status of a city trunk road intersection shown in the figure comprises the following steps:

[0101] S1. Collect historical vehicle data of the target intersection, preprocess the data, and calculate the historical discrimination threshold of the oversaturation state;

[0102] S2, collecting real-time vehicle data of each cycle of the target intersection, preprocessing the data, and judging whether it belongs to an oversaturation cycle based on the oversaturation state historical discrimination threshold;

[0103] S3. If it is an oversaturation period, determine whether the target intersection is overloaded during the oversaturation period;

[0104] S4. If the target intersection is not overloaded, determine whether the target intersection is unbalanced within the oversaturation period;

[0105] S5. Output the judgment result.

[0106] In a more preferred embodiment, the data preprocessing in steps S1 and S2 includes data reduction, data conversion, and data cleaning performed in sequence;

[0107] In this embodiment, the data reduction method includes: retaining the license plate data, the vehicle entrance lane data, and the vehicle lane data in the collected vehicle data, and deleting the remaining data;

[0108] In this embodiment, the method of data conversion includes: assigning lane directions to all lanes of each entrance road of the target intersection, and superimposing the lane directions to the collected vehicle data;

[0109] The data cleaning includes executing in sequence: cleaning duplicate data and eliminating abnormal data.

[0110] In this embodiment, the method for cleaning up duplicate data includes:

[0111] Extract all license plate data of an entrance lane of the target intersection, group the extracted license plate data by license plate type, and sort the license plate data in each group by collection time;

[0112] Divide each group of sorted license plate data into a number of first time intervals according to the first time interval, and determine whether there are duplicate license plates in each first time interval: if so, retain the last vehicle data of the duplicate license plate sorted by time, and delete the remaining vehicle data of the duplicate license plate;

[0113] Extract all license plate data of the next entrance road of the target intersection until all entrance roads of the target intersection have been cleared of duplicate data;

[0114] The method for removing abnormal data includes: deleting license plate data with missing license plate recognition and deleting license plate data with license plate character recognition errors.

[0115] Preferably, the first time interval is 100 seconds.

[0116] Embodiment 2:

[0117] A method for determining the traffic state of an urban trunk intersection, based on Example 1, a method for calculating a historical determination threshold value of an oversaturation state includes:

[0118] S201, dividing the preprocessed historical vehicle data into a plurality of second time intervals according to a second time interval;

[0119] S203, calculating the hourly flow rate and equivalent density of the target intersection in the most recent second time interval, and establishing an hourly flow rate-equivalent density map;

[0120] S204: Based on the hourly flow rate-equivalent density diagram, determine whether there is an oversaturation state in the second time interval:

[0121] If so, obtaining a supersaturation state historical discrimination threshold based on the hourly flow rate-equivalent density diagram in the second time interval;

[0122] If not, the most recent second time interval is replaced by the previous second time interval, and the process is iterated forward until the oversaturation state historical judgment threshold is obtained.

[0123] In this embodiment, the equivalent density is calculated by the following formula: Where: D is equivalent density; hV is hourly flow rate; L is average section length; T is average travel time of section;

[0124] In this embodiment, the method for establishing the hourly flow rate-equivalent density map includes:

[0125] S2031, establishing a data set based on the obtained hourly flow rate and equivalent density, wherein the data set includes a number of scattered points;

[0126] S2032. For all scattered points, take 5 as the sliding window and 1 as the step size, and perform sliding average to obtain a number of scattered points after sliding.

[0127] S2033, establishing a two-dimensional coordinate system with equivalent density as the horizontal coordinate and hourly flow rate as the vertical coordinate; marking all the scattered points after sliding in the two-dimensional coordinate system.

[0128] The method for determining whether an oversaturation state exists within the second time interval includes:

[0129] S2041. Perform linear fitting of a single segment straight line on all the scattered points after sliding in the two-dimensional coordinate system. The fitting result is as follows: Figure 2 As shown, Figure 2 The horizontal axis is the equivalent density and the vertical axis is the hourly flow rate; the first root mean square error is calculated based on the fitting results;

[0130] S2042. Perform piecewise linear fitting of two straight lines on all the scattered points after sliding in the two-dimensional coordinate system. The fitting results are as follows: Figure 3 As shown, Figure 3 The horizontal axis is the equivalent density and the vertical axis is the hourly flow rate; the second root mean square error is calculated based on the fitting results;

[0131] S2043, comparing the first root mean square error with the second root mean square error: if the first root mean square error is greater than the second root mean square error, an oversaturation state exists in the corresponding second time interval; if the first root mean square error is less than or equal to the second root mean square error, an oversaturation state does not exist in the corresponding second time interval. That is, if the optimal fitting solution is two straight line segmented fitting, it is judged that an oversaturation state exists; otherwise, if the optimal fitting solution is a single straight line fitting, an oversaturation state does not exist.

[0132] More preferably, all segmentation methods are traversed in the two-dimensional coordinate system, that is, various situations in which the scattered points after sliding are divided into two parts are traversed, and linear fitting is performed on the scattered points after sliding of the two parts respectively: one part adopts normal linear fitting based on the least square method, and the other part adopts horizontal line fitting with a fixed slope of 0 and the minimum RMSE as the goal. The fitting scheme with the smallest root mean square error among all traversal schemes is taken as the second root mean square error.

[0133] In this embodiment, the method for obtaining the oversaturation state historical discrimination threshold is: defining the smoothing density corresponding to the intersection point of two straight lines in the result of the piecewise linear fitting as the critical density; and calculating the oversaturation state historical discrimination threshold = 1.2×critical density.

[0134] In this embodiment, the average section length L refers to the average length of each entrance road, and the average section travel time T refers to the average time taken by a vehicle from entering the entrance road to passing the stop line.

[0135] Preferably, the second time interval is in natural days, and the second time interval is 15 minutes; Figure 4 As shown, Figure 4 The middle horizontal axis is the cycle number, the left vertical axis is the cycle flow, and the right vertical axis is the average travel time.

[0136] In this embodiment, when performing real-time analysis, for the same target intersection at the end of each day, if there is an oversaturation state on that day, the historical discrimination threshold is updated with the discrimination threshold of that day; if there is no oversaturation phenomenon on that day, the historical discrimination threshold is not updated. The updated historical discrimination threshold can serve the real-time discrimination of the oversaturation state on the next day.

[0137] Embodiment 3:

[0138] A method for determining the traffic state of an urban trunk intersection, based on Embodiment 1 or 2, a method for determining whether it belongs to an oversaturation period based on an oversaturation state historical determination threshold is as follows:

[0139] S205, based on the pre-processed real-time vehicle data of the target intersection in each cycle, the equivalent density in the cycle is calculated in real time, and the specific calculation process can adopt the same calculation method and formula as the equivalent density in Example 2;

[0140] S206, comparing the equivalent density in the period with the historical threshold for judging the oversaturation state:

[0141] If the equivalent density in the cycle is greater than or equal to the historical discrimination threshold of the oversaturation state, the cycle belongs to the oversaturation cycle;

[0142] If the equivalent density in this period is less than the historical judgment threshold of the oversaturation state, then this period does not belong to the oversaturation period.

[0143] In a more preferred embodiment, the method for determining whether the target intersection is overloaded includes:

[0144] S301, calculating the saturation flow rate of the target intersection and the cycle flow of each cycle;

[0145] S302, respectively calculating the flow ratio of each entrance of the target intersection: In the formula, y i is the flow ratio of the i-th entrance of the target intersection; q i is the cycle flow of the i-th entrance of the target intersection; s i is the saturated flow rate of the i-th entrance of the target intersection;

[0146] S303, calculate the sum Y of the flow ratios of all entrances of the target intersection: Y = ∑y i ;

[0147] S304. Calculate the overload coefficient η: Where: λ is the green-to-signal ratio of the cycle; G is the total effective green light time in the current cycle; C is the cycle length;

[0148] S305. Determine η: If η>1, the target intersection is overloaded.

[0149] Among them, in the process of calculating the periodic flow of each cycle at the target intersection, by extracting the vehicle model in the vehicle data, the standard flow of different vehicle models is converted according to the following formula: Pcu=α×Veh; where; Pcu is the standard vehicle flow; Veh is the unit vehicle flow; α is the vehicle model coefficient.

[0150] In this embodiment:

[0151] If the vehicle type is a light truck, take α = 1.2;

[0152] If the vehicle type is a medium-sized ordinary passenger car or medium-sized ordinary truck, take α = 2.4;

[0153] If the vehicle type is a large ordinary passenger car or a heavy ordinary truck, take α = 3.6;

[0154] If the vehicle type is other, take α=1.

[0155] When there are three or more consecutive cycles in the overload state, it can be considered that the target intersection has a systematic overload problem.

[0156] For the overload phenomenon, the method of simply increasing the cycle length can be used to alleviate it, but the effect is usually limited, and the marginal benefits tend to decrease with the increase of the cycle length. Therefore, in this embodiment, various means of improving the traffic capacity of the intersection can be considered, such as optimizing the signal phase scheme, dividing the functions of the entrance lanes, separating traffic flow interference (such as pedestrians and non-motorized vehicles), and increasing the number of entrance lanes. In addition, the causes of intersection overload are mostly more macroscopic road network layout and supply and demand factors. When it is difficult to resolve a single node through traffic design means, traffic flow organization means (such as one-way, no left, diversion, upstream flow control) and traffic demand management measures should be comprehensively considered.

[0157] Embodiment 4:

[0158] A method for determining the traffic state of an urban trunk intersection. Based on any of the above embodiments, in this embodiment, the saturation flow rate s of the i-th entrance of the target intersection is i Calculated by the following method:

[0159] S3021. Based on the historical vehicle data of the target intersection, identify at least 50 saturated vehicle fleets of different cycles;

[0160] S3022, obtaining a set of headway time sequences in each saturated fleet;

[0161] S3023, eliminating the first four headway times in each group of headway time sequences, and obtaining the mean headway time of the remaining groups;

[0162] S3024. Calculate the average of all the remaining headway time averages of the single group to obtain the average saturated headway time, and calculate the saturated flow rate based on the average saturated headway time.

[0163] Among them, the identification method of saturated fleet is as follows: Figure 5 As shown, it can be specifically described as:

[0164] S30211. Identify the leading vehicle in the convoy: extract the vehicle whose headway is greater than t+20s and whose headway of the eight vehicles that pass in succession is less than or equal to 6s as the leading vehicle in the convoy; where t is the red light duration of the lane corresponding to the entrance lane;

[0165] S30212. Starting from the head vehicle of the convoy, all vehicles whose headway time is continuously less than or equal to 6 seconds are regarded as a saturated convoy.

[0166] It should be noted that Figure 5The "CBV" here stands for the lead vehicle of the convoy, and the "CBV convoy" stands for the saturated convoy.

[0167] In this embodiment, the method for determining whether the target intersection is unbalanced includes:

[0168] Calculate the flow ratio of each lane in each entrance lane of the target intersection: In the formula, y im is the flow ratio of the mth lane in the i-th entrance of the target intersection; q im is the cycle flow of the mth lane in the i-th entrance of the target intersection; s im is the saturation flow rate of the mth lane in the i-th entrance of the target intersection;

[0169] Calculate the green signal ratio λ of the mth lane in the i-th entrance lane im : In the formula, g im is the effective green light time corresponding to the mth lane in the i-th entrance lane; C is the cycle length;

[0170] Judgment: If any y im >λ im , then the target intersection is unbalanced.

[0171] The problem of intersection imbalance can be solved by re-optimizing the signal timing plan.

[0172] To sum up, under the demand of implementing refined traffic control, this application can fully tap into the existing data resources, develop a traffic status recognition algorithm based on license plate data, and realize the ability to describe traffic status with high temporal and spatial resolution on a cycle-by-cycle, lane-level, and quasi-real-time basis. On this basis, it can also conduct a detailed diagnosis of problems such as overload and imbalance at signalized intersections, and form an optimized signal timing plan, which can realize the value conversion of existing data resources, improve the operating efficiency of signalized intersections, and enhance the level of refined road traffic management.

[0173] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0174] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

Claims

1. A method for determining the traffic status of an urban trunk road intersection, characterized in that: The following steps are involved: Collect historical vehicle data of the target intersection, preprocess the data, and calculate the historical threshold for oversaturation status; Collect real-time vehicle data of each cycle at the target intersection, pre-process the data, and determine whether it belongs to an oversaturation cycle based on the oversaturation state historical discrimination threshold; If it is an oversaturation period, then during the oversaturation period, determine whether the target intersection is overloaded; If the target intersection is not overloaded, then during the oversaturation period, it is determined whether the target intersection is unbalanced; Output the judgment result; The method for calculating the oversaturation state history discrimination threshold comprises: dividing the preprocessed historical vehicle data into a plurality of second time intervals according to a second time interval; In the most recent second time interval, the hourly flow rate and equivalent density of the target intersection are calculated, and an hourly flow rate-equivalent density map is established; Based on the hourly flow rate-equivalent density diagram, determine whether there is an oversaturation state in the second time interval: If so, obtaining a supersaturation state historical discrimination threshold based on the hourly flow rate-equivalent density diagram in the second time interval; If not, replace the most recent second time interval with the previous second time interval, and iterate forward until the oversaturation state historical discrimination threshold is obtained; The equivalent density is calculated by the following formula: Where: D is equivalent density; hV is hourly flow rate; L is average section length; T is average travel time of section; The method for establishing an hourly flow rate-equivalent density diagram comprises: Establishing a data set with the obtained hourly flow rate and equivalent density, wherein the data set includes a plurality of scattered points; For all scattered points, perform sliding average with a specified sliding window and a specified step size to obtain a number of scattered points after sliding; A two-dimensional coordinate system is established with equivalent density as the horizontal coordinate and hourly flow rate as the vertical coordinate; all the scattered points after sliding are marked in the two-dimensional coordinate system; The method for determining whether an oversaturation state exists within the second time interval includes: Perform linear fitting of a single segment straight line on all the scattered points after sliding in the two-dimensional coordinate system, and calculate the first root mean square error based on the fitting result; Perform piecewise linear fitting of two straight lines on all the scattered points after sliding in the two-dimensional coordinate system, and calculate the second root mean square error based on the fitting results; Compare the first RMS error to the second RMS error: If the first root mean square error is greater than the second root mean square error, an oversaturation state exists in the corresponding second time interval; If the first root mean square error is less than or equal to the second root mean square error, then there is no oversaturation state in the corresponding second time interval; The method for obtaining the oversaturation state historical discrimination threshold is as follows: in the result of the piecewise linear fitting, the smooth density corresponding to the intersection point of two straight lines is defined as the critical density; the oversaturation state historical discrimination threshold is calculated as 1.2×critical density; The method for determining whether the target intersection is overloaded comprises: Calculate the saturated flow rate and cycle flow of each cycle at the target intersection; the saturated flow rate is estimated in advance, and the cycle flow is calculated in real time. Real-time calculation means that after each cycle, the real-time vehicle data of each cycle is collected and calculated; Calculate the flow ratio of each entrance of the target intersection separately: In the formula, y i is the flow ratio of the i-th entrance of the target intersection; q i is the cycle flow of the i-th entrance of the target intersection; s i is the saturated flow rate of the i-th entrance of the target intersection; Calculate the sum of the flow ratios of all entrances at the target intersection: Y = ∑y i ; Calculate the overload factor η: Where: λ is the green-to-signal ratio of the cycle; G is the total effective green light time in the current cycle; C is the cycle length; Determine η: If η>1, the target intersection is overloaded; The method for determining whether the target intersection is unbalanced comprises: Calculate the flow ratio of each lane in each entrance lane of the target intersection: In the formula, y im is the flow ratio of the mth lane in the i-th entrance of the target intersection; q im is the cycle flow of the mth lane in the i-th entrance of the target intersection; s im is the saturation flow rate of the mth lane in the i-th entrance of the target intersection; Calculate the green signal ratio λ of the mth lane in the i-th entrance lane im : In the formula, g im is the effective green light time corresponding to the mth lane in the i-th entrance lane; C is the cycle length; Judgment: If any y im >λ im , then the target intersection is unbalanced.

2. A method for determining the traffic status of an urban trunk road intersection according to claim 1, characterized in that: The data preprocessing includes data reduction, data conversion, and data cleaning performed in sequence; The data reduction method includes: retaining the license plate data, the vehicle entrance lane data, and the vehicle lane data in the collected vehicle data, and deleting the remaining data; The data conversion method includes: assigning lane directions to all lanes of each entrance road of the target intersection, and superimposing the lane directions onto the collected vehicle data; The data cleaning includes executing in sequence: cleaning duplicate data and eliminating abnormal data.

3. A method for determining the traffic status of an urban trunk intersection according to claim 2, characterized in that: The method for cleaning duplicate data includes: Extract all license plate data of an entrance lane of the target intersection, group the extracted license plate data by license plate type, and sort the license plate data in each group by collection time; Divide each group of sorted license plate data into a number of first time intervals according to the first time interval, and determine whether there are duplicate license plates in each first time interval: if so, retain the last vehicle data of the duplicate license plate sorted by time, and delete the remaining vehicle data of the duplicate license plate; Extract all license plate data of the next entrance road of the target intersection until all entrance roads of the target intersection have been cleared of duplicate data; The method for removing abnormal data includes: deleting license plate data with missing license plate recognition and deleting license plate data with license plate character recognition errors.

4. A method for determining the traffic status of an urban trunk intersection according to claim 1, characterized in that: Methods for judging whether it is an oversaturation period based on the oversaturation state history judgment threshold include: Based on the pre-processed real-time vehicle data of each cycle of the target intersection, the equivalent density within the cycle is calculated in real time; Compare the equivalent density within this period with the historical threshold for supersaturation: If the equivalent density in the cycle is greater than or equal to the historical discrimination threshold of the oversaturation state, the cycle belongs to the oversaturation cycle; If the equivalent density in this period is less than the historical judgment threshold of the oversaturation state, then this period does not belong to the oversaturation period.

5. A method for determining the traffic status of an urban trunk intersection according to claim 1, characterized in that: In the process of calculating the cycle flow of each cycle at the target intersection, by extracting the vehicle model in the vehicle data, the standard flow conversion of different vehicle models is performed according to the following formula: Pcu = α × Veh; where Pcu is the standard vehicle flow; Veh is the unit vehicle flow; α is the vehicle model coefficient; The saturation flow rate of the i-th entrance of the target intersection is calculated as follows: Based on the historical vehicle data of the target intersection, identify at least N saturated fleets with different periods, where N ≥ 50; Get a set of headway time sequences in each saturated fleet; Eliminate the first four headway times in each group of headway time sequences to obtain the mean headway time of the remaining groups; The average of all the single-group remaining headway means is calculated to obtain the average saturated headway, and the saturated flow rate is calculated based on the average saturated headway.

6. A method for determining the traffic status of an urban trunk road intersection according to claim 5, characterized in that: The saturated fleet is identified by: Identify the leading vehicle in a convoy: extract the vehicle whose headway is greater than t+20s and whose headway of the eight vehicles that pass in succession is less than or equal to 6s as the leading vehicle in the convoy; where t is the red light duration of the corresponding lane of the entrance lane; Starting from the head vehicle of the convoy, all vehicles whose headway time is continuously less than or equal to 6s are regarded as a saturated convoy.

Citation Information

Patent Citations

  • Road traffic condition evaluation method

    CN105825683A

  • Road network tide traffic flow variable guide lane control method

    CN106548633A