An intelligent traffic signal parameter configuration method and system

Abstract

The present invention relates to the technical field of traffic control, and specifically to an intelligent traffic signal parameter configuration method and system, which includes the following steps: obtaining vehicle passing behaviors, extracting the lateral offset and longitudinal delay time series before and after the turning action, identifying the signal behavior delay section, extracting the waiting duration and release completion time point of the vehicle at the end of the lane, calculating the time gap, detecting the longitudinal distance of the queued vehicles in the real-time cycle, updating the control cycle of the end section, and obtaining the regulation result of the green light duration at the end section. In the present invention, through the detailed identification and analysis of vehicle behaviors and position characteristics, the signal scheduling is optimized, the signal light cycle is dynamically adjusted through real-time data, traffic congestion caused by untimely response is reduced, and for the calculation of the waiting duration and release time of the vehicle at the end, a variable green light time period is allocated for each lane, so that the green light time at the end section can be flexibly regulated according to actual needs, further reducing traffic delays and improving road utilization efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic control, and particularly to an intelligent traffic signal parameter configuration method and system. Background Art

[0002] The technical field of traffic control includes a series of methods and devices for solving traffic management and optimization problems. The core content is to improve road usage efficiency, reduce traffic congestion, and enhance traffic safety. Specifically, it involves the design, implementation, adjustment, and monitoring technologies of traffic signals, including signal control systems, vehicle detection systems, and traffic data analysis methods. This field also includes the study of traffic flow patterns and strategies for adjusting traffic flow at different times and locations through technical means.

[0003] Among them, the intelligent traffic signal parameter configuration method refers to the process of intelligently setting and adjusting the control parameters of traffic signals. This technical topic mainly involves how to adjust the green light duration and red light duration of traffic signals according to real-time traffic flow data to optimize traffic mobility and reduce traffic congestion. This method includes collecting traffic data, analyzing traffic data, and adjusting the parameters of signal operation according to the analysis results. This configuration method relies on traffic flow detection devices and data processing software to update the operation mode of traffic signals in real time to adapt to changing traffic conditions.

[0004] Existing traffic control technologies mainly rely on fixed signal cycles and static signal control modes, which cannot fully adapt to real-time road conditions in a dynamically changing traffic environment. Especially during traffic peaks or emergencies, fixed signal settings are difficult to cope with instantaneous traffic flow increases or decreases, easily causing traffic congestion and resource waste. Existing methods also show limitations in the recognition and analysis of vehicle behaviors, lacking a detailed judgment of the impact of specific behaviors. For example, the recognition of lateral offset and longitudinal delay is insufficient, resulting in signal control being unable to accurately reflect the actual passing state of vehicles and affecting the smoothness of traffic flow. These problems not only reduce the efficiency of traffic signals but also increase the risks of road usage. Especially in the management of complex intersections, the lack of an effective adjustment mechanism can lead to traffic paralysis and safety accidents. Summary of the Invention

[0005] In order to solve the problems existing in the prior art that in a dynamically changing traffic environment, it is impossible to fully adapt to real-time road condition changes. Especially during traffic peaks or emergencies, fixed signal settings are difficult to cope with instantaneous traffic flow increases or decreases, easily causing traffic congestion and resource waste. Existing methods also show limitations in the recognition and analysis of vehicle behaviors, lacking a detailed judgment of the impacts of specific behaviors. For example, the recognition of lateral offset and longitudinal delay is insufficient, resulting in the signal control being unable to accurately reflect the actual passing state of vehicles and affecting the smoothness of traffic flow. The problems not only reduce the efficiency of traffic signals but also increase the risks of road use. Especially in the management of complex intersections, the lack of an effective adjustment mechanism leads to traffic paralysis and safety accidents. The present invention provides an intelligent traffic signal parameter configuration method and system. The technical solution is as follows:

[0006] On the one hand, an intelligent traffic signal parameter configuration method is provided. The method includes:

[0007] S1: Obtain the vehicle passing data of the lanes at the signal-controlled intersection, extract the lateral offset positions and longitudinal delay time series before and after the turning actions, and judge through the vehicle deceleration frequency corresponding to the behavior to generate an interference behavior response tag group;

[0008] S2: Call the marked positions in the interference behavior response tag group, obtain the starting time of the green light, the starting time of the leading vehicle, and the releasing time of the trailing vehicle within the section, identify the behaviors and passing deviation segments within the signal control cycle, and generate a signal behavior delay section;

[0009] S3: According to the channels indicated in the signal behavior delay section, extract the waiting duration of the trailing vehicles at the end of the green light and the releasing completion time points within two cycles of the channels, calculate the time gap with the real-time green light end point, judge whether the channel release is incomplete, screen the waiting durations corresponding to the incomplete channels as the priority sorting basis, and generate a channel green light variable time period configuration table;

[0010] S4: Call the starting configuration values in the channel green light variable time period configuration table, detect the longitudinal distance of the queuing vehicles in the real-time cycle, and mark the continuous acceleration time points of three adjacent vehicles. If the acceleration occurrence times are all lower than the signal preparation switching point, record the synchronous start state and generate a signal cycle start instruction.

[0011] As a further solution of the present invention, the interference behavior response tag group includes a traffic interruption identification number, an interference behavior classification code, and a lane space mapping position. The signal behavior delay section includes a signal start response delay length, a vehicle release rhythm lag length, and a behavior trigger and traffic rhythm offset length. The channel green light variable time period configuration table includes a channel sequence sorting weight, a green light time period adjustment range, and a scheduling control effective time. The signal cycle start instruction includes a cycle switching trigger state, a synchronous start trigger moment, and a vehicle distance compression recognition state.

[0012] As a further solution of the present invention, the obtaining steps of the interference behavior response tag group are specifically as follows:

[0013] S101: Obtain the vehicle traffic data of the lanes at the signal-controlled intersection, extract the lateral coordinate change amount and the longitudinal position change time sequence before and after each vehicle executes a turning action, and calculate the lateral offset distance and the longitudinal action delay period based on the lateral coordinate difference and the time length of the longitudinal sequence each time, and generate an offset and delay time sequence parameter set;

[0014] S102: According to the lateral offset distance and the longitudinal action delay period of the vehicle in the offset and delay time sequence parameter set, and in combination with the speed change information of the vehicle at consecutive moments before and after the turning action, count the deceleration times of each vehicle in the offset and delay stages, identify the synchronous distribution relationship between the deceleration times and the delay periods, and generate the deceleration frequency of the vehicle turning stage;

[0015] S103: Based on the deceleration frequency of the vehicle turning stage, judge the vehicle behavior whose deceleration frequency exceeds the preset turning behavior critical frequency, bind the corresponding behavior action to the position of the road section where it is located, and generate an interference behavior response tag group.

[0016] As a further solution of the present invention, the obtaining steps of the signal behavior delay section are specifically as follows:

[0017] S201: Call the road position section identified in the interference behavior response tag group, obtain three time data of the green light start time, the first vehicle start time, and the last vehicle release time in the corresponding section during the signal control cycle, calculate the time intervals between the first vehicle response time and the last vehicle traffic time respectively, and generate traffic key time interval data;

[0018] S202: Based on the time intervals between the first vehicle response time and the last vehicle traffic time values in the traffic key time interval data, call the total number of vehicles and the vehicle position distribution in the corresponding section during the green light cycle, identify the total queue length and the queue sparsity level, and obtain the traffic density distribution information;

[0019] S203: According to the traffic density distribution information, combined with the time interval distribution of the front, middle, and rear position points in the queue during the key traffic time period, calculate the traffic density offset eigenvalue, identify the position interval of the density anomaly section in the green light control cycle, and generate the traffic density offset time period.

[0020] The formula for calculating the traffic density offset eigenvalue is as follows:

[0021] ;

[0022] Where, AD represents the traffic density offset eigenvalue, n represents the number of density detection sections participating in the evaluation within the green light cycle, represents the vehicle queue length within the i-th detection section, represents the passing time length of the tail vehicle within the i-th detection section, represents the average longitudinal response density value of the front, middle, and rear groups of vehicles in the i-th detection section, represents the delay time of the tail vehicle in the queue section responding to the head vehicle in the i-th detection section, and BD represents the set reference traffic density benchmark value;

[0023] S204: Based on the corresponding section positions in the traffic density offset time period, evaluate the correlation between the time period within the signal control cycle and the vehicle release behavior segment, screen out the continuous segments with offset vehicle release behavior, mark them as the areas where traffic behavior delays occur, and generate the signal behavior delay section.

[0024] As a further solution of the present invention, the steps for obtaining the channel green light variable time period configuration table are specifically as follows:

[0025] S301: According to the channel number indicated in the signal behavior delay section, extract the stationary duration and release completion time point of the tail vehicle at the end of the green light within two consecutive signal cycles of the channel, and combine the time difference between the release completion time point and the end time point of the green light in the current cycle to obtain the release lag information of the channel tail section;

[0026] S302: Based on the release lag information of the channel tail section, determine whether the release lag value exceeds the channel release completion benchmark threshold, screen out the channel numbers with uncompleted releases, extract the waiting time of the tail vehicles in the corresponding channels, sort them according to the waiting duration of the uncompleted channels, and generate the channel release waiting priority sequence;

[0027] S303: Call the sorting information in the channel release waiting priority sequence, sequentially configure additional green light time periods for the channels, adjust the length of the channel green light time period within the total green light duration, record the channel number and the corresponding adjusted green light duration, and generate the channel green light variable time period configuration table.

[0028] As a further solution of the present invention, the step of obtaining the signal cycle start instruction is specifically as follows:

[0029] S401: Call the channel start configuration time value recorded in the channel green light variable time period configuration table, detect the longitudinal distance distribution of vehicles in the corresponding channel queuing area within the real-time signal cycle, identify the continuous acceleration start time points of three adjacent vehicles in the queuing section, and generate the queuing acceleration start time;

[0030] S402: According to the time points of the acceleration actions of three adjacent vehicles in the queuing acceleration start time, judge whether the acceleration times are all earlier than the signal preparation switching time point within the channel green light variable time period. If the judgment condition is met, mark it as the synchronous start state, bind the channel number and status information, and generate the signal cycle start instruction.

[0031] As a further solution of the present invention, the formula for judging whether the acceleration time is earlier than the signal preparation switching time point within the channel green light variable time period is as follows:

[0032] ;

[0033] Where, AT represents the average start timing difference, , , respectively represent the acceleration start times of the first, second, and third vehicles in the group, represents the speed at the sampling moment before the a-th vehicle's acceleration action, OV represents the passing reference speed set for the road section, k is the weighting coefficient, and m is the number of vehicles.

[0034] As a further solution of the present invention, the method further includes step S5:

[0035] S5: Call the marked period of the signal cycle start instruction, screen the passing trajectories of the trailing vehicle in the latter part of the green light, compare the vehicle speed change trend with the duration of the latter part. If the speed remains rising and does not reach stability, calculate the time required to reach the predetermined release point and update the control cycle of the latter part to obtain the continuous regulation result of the latter part of the green light;

[0036] The continuous regulation result of the latter part of the green light includes the remaining release path continuation duration, the predicted period for the trailing vehicle to complete release, and the signal holding recommended time window.

[0037] As a further solution of the present invention, the step of obtaining the continuous regulation result of the latter part of the green light is specifically as follows:

[0038] S501: Call the signal cycle number marked in the signal cycle start instruction, screen the trajectory data of the trailing vehicle in the latter part of the green light under the corresponding cycle, extract the continuous speed sequence and timestamp data from the start moment of the latter part to the moment when the vehicle exits the release line, and generate the passing trajectory sequence of the latter part;

[0039] S502: Analyze the speed change trend of the trailing vehicle within the green light duration of the tail section based on the speed sequence in the tail section passing trajectory sequence, extract the speed increase value at the end of the sequence, and compare it with the corresponding duration of the tail section. If the speed continues to rise and does not enter the stable range, calculate the additional time required for the vehicle to reach the predetermined release point, and generate the remaining release time interval for the tail section.

[0040] S503: Invoke the additional time value required for the channel in the remaining release time interval of the tail section, update the real-time green light control period of the tail section, correct the original configured end time point of the green light of the tail section, and reset the control interval of the tail section of the channel to obtain the continuous regulation result of the green light of the tail section.

[0041] On the other hand, the intelligent traffic signal parameter configuration system is used to execute the above intelligent traffic signal parameter configuration method, and the system includes:

[0042] The behavior data recognition module detects the vehicle passing behavior of the lanes at the signal-controlled intersection, collects the lateral offset position and longitudinal delay time series of the vehicle, counts the deceleration frequency of the vehicle, and generates an interference behavior response tag group.

[0043] The interference behavior analysis module, based on the interference behavior response tag group, identifies the key behavior actions that cause traffic interruption, binds the key behavior actions to the location, and generates a signal behavior delay section.

[0044] The signal delay calculation module invokes the signal behavior delay section, obtains the start time of the green light, the starting time of the leading vehicle, and the release time of the trailing vehicle within the section, calculates the time interval, and generates a channel green light variable time period configuration table.

[0045] The green light time dynamic configuration module, based on the channel green light variable time period configuration table, analyzes the completion situation of the channel release, calculates the waiting time of the uncompleted channel, reallocates the green light time period as needed, and generates a signal cycle start instruction.

[0046] The signal start module invokes the signal cycle start instruction, detects and evaluates the longitudinal distance and continuous acceleration points of the queued vehicles, adjusts the signal start time according to the vehicle speed change trend and the tail section duration, and obtains the continuous regulation result of the green light of the tail section.

[0047] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0048] By carefully identifying and analyzing the vehicle behavior and position characteristics, the adaptability and accuracy of signal control are effectively improved. The lateral offset and longitudinal delay behaviors of the vehicle are extracted and combined with the actual traffic interruption behavior, enabling the traffic signal to more accurately reflect the real-time road conditions and reducing traffic jams caused by signal inadaptability. During the identification process, the analysis of the vehicle deceleration frequency further optimizes the signal scheduling. By dynamically adjusting the signal light cycle through real-time data, the traffic congestion caused by untimely response is reduced. The calculation of the waiting time and release time of the trailing vehicle, and the variable green light time period allocated for each lane, provide a more reasonable prediction of the passing time, ensuring the continuity and high efficiency of the traffic flow. By integrating the vehicle speed change trend and the adjustment of the control cycle, the green light time at the tail section can be flexibly regulated according to the actual needs, further reducing traffic delays and improving the road utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the working process of the present invention;

[0050] Figure 2 It is a system flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] The following describes the technical solutions in the present invention with reference to the accompanying drawings.

[0052] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as an "example" in the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, the use of the word "example" is intended to present concepts in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one of the two.

[0053] To make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0054] Please refer to Figure 1 , the embodiments of the present invention provide an intelligent traffic signal parameter configuration method, and the processing flow of this method may include the following steps:

[0055] S1: Obtain the vehicle passing data of the lanes at the signal control intersection, extract the lateral offset position and longitudinal delay time series before and after the steering action, judge through the vehicle deceleration frequency corresponding to the behavior, identify the behavior actions causing traffic interruption and bind them to the location, and generate an interference behavior response tag group;

[0056] S2: Call the marked positions in the interference behavior response tag group to obtain the starting time of the green light, the starting time of the first vehicle's start, and the release time of the last vehicle within the section, calculate the time intervals among the three, compare with the queue length, identify the behaviors and passing deviation segments within the signal control period, and generate a signal behavior delay section;

[0057] S3: According to the channels indicated in the signal behavior delay section, extract the waiting duration of the tail vehicle at the end of the green light and the release completion time point within two cycles of the channel, calculate the time gap with the real-time green light end point, determine whether the channel release is incomplete, select the waiting duration corresponding to the incomplete channel as the priority sorting benchmark, and allocate variable green light time periods to each channel in sequence to generate a channel green light variable time period configuration table;

[0058] S4: Call the starting configuration value in the channel green light variable time period configuration table, detect the longitudinal distance of the queued vehicles in the real-time cycle, and mark the continuous acceleration time points of three adjacent vehicles. If the acceleration occurrence times are all lower than the signal preparation switching point, record the synchronous start state and generate a signal cycle start instruction;

[0059] S5: Call the marked cycle of the signal cycle start instruction, screen the passing trajectories of the tail vehicles in the latter part of the green light, compare the vehicle speed change trend with the duration of the tail section. If the speed keeps rising and does not reach stability, calculate the time required to reach the predetermined release point and update the tail section control cycle to obtain the tail section green light continuous regulation result;

[0060] The interference behavior response tag group includes a traffic interruption identification number, an interference behavior classification code, and a lane space mapping position. The signal behavior delay section includes a signal start response delay length, a vehicle release rhythm lag length, and a behavior trigger and passing rhythm offset length. The channel green light variable time period configuration table includes a channel sequence sorting weight, a green light time period adjustment range, and a scheduling control effective time. The signal cycle start instruction includes a cycle switching trigger state, a synchronous start trigger moment, and a vehicle distance compression identification state. The tail section green light continuous regulation result includes a remaining release path continuation duration, a last vehicle release completion prediction period, and a signal holding recommended time window.

[0061] The specific steps for obtaining the interference behavior response tag group are as follows:

[0062] S101: Obtain the vehicle passing data of the lanes at the signal control intersection, extract the lateral coordinate change amount and the longitudinal position change time sequence before and after each vehicle executes a turning action, and calculate the lateral offset distance and the longitudinal action delay cycle based on each lateral coordinate difference and the time length of the longitudinal sequence to generate an offset and delay time sequence parameter set;

[0063] It is necessary to rely on the camera recognition equipment and geomagnetic induction equipment installed at signal-controlled intersections to record the vehicles passing through the lanes frame by frame all day long, and based on the positions of each vehicle in the frame images at different time points, identify its trajectory direction. Combining the image frame sequences of each vehicle before and after the turning action, extract the horizontal coordinate values of each vehicle within a continuous time period, identify the coordinate differences before and after its turning, and at the same time obtain the corresponding time range of this process. Through the horizontal coordinate point sequence, the offset direction and amplitude of the vehicle trajectory in the lane can be marked. Set in an actual crossroads, after the signal is released, a car deviates from the straight lane to the left-turning direction. The front of the car shows an obvious leftward movement trend at the 15th frame and completely enters the left-turn lane at the 22nd frame. This time period is marked as the turning section. At the same time, extract the time sequence of the vehicle's longitudinal displacement, combine the geomagnetic data to determine its relative coordinate sequence, generate the horizontal offset data and longitudinal action duration sequence of each vehicle. After completing the annotation of this process for the vehicle, combine the lane number, passing time period and vehicle ID to establish a data set with a multi-field structure, form the statistical parameters of offset and delay, and form the offset and delay time sequence parameter set.

[0064] S102: According to the vehicle horizontal offset distance and longitudinal action delay period in the offset and delay time sequence parameter set, combine the speed change information of the vehicle at consecutive moments before and after the turning action, count the deceleration times of each vehicle in the offset and delay stages, identify the synchronous distribution relationship between the deceleration times and the delay periods, and generate the deceleration frequency of the vehicle turning stage;

[0065] It is necessary to further call the speed change sequence of the vehicle per second before and after turning, synchronously extract the speed values of each time period in combination with the frame rate, and judge whether there is a continuous deceleration process. Taking the left-turn lane in the intersection as an example, during the turning process of a bus, its speed drops from the original 9 meters per second to 7 meters, and then to 6 meters, indicating that this process is affected by the turning path or the front block, and it needs to be marked as an effective deceleration action. At the same time, detect whether the vehicle has experienced a speed drop twice or more in the time period before and after the offset, record the deceleration action and its occurrence time period, and count the occurrence times. After making this judgment on the vehicle, construct a result list containing the deceleration action times of each vehicle, pair the deceleration times with their corresponding longitudinal delay times, and judge whether they show a synchronous distribution structure, that is, whether most deceleration actions occur in the horizontal offset section or its time period. If it is satisfied, it is classified into the synchronous group. In the crossroads test, the vehicles in multiple lanes show an obvious synchronous deceleration trend, especially when the vehicles are close to the signal light area during the peak period, this kind of data can be directly recorded and summarized as the behavioral performance characteristics of the corresponding lane in the turning stage, and the deceleration frequency of the vehicle turning stage is generated.

[0066] S103: Based on the deceleration frequency during the vehicle steering phase, determine the vehicle behavior with a deceleration frequency exceeding the preset critical frequency of steering behavior, bind the corresponding behavioral actions to the position of the road section where they occur, and generate a set of interference behavior response tags;

[0067] It is necessary to set the judgment conditions for steering behavior, screen and process the deceleration frequency data, and clarify that some vehicle behaviors are abnormal. Taking Area A of the crossroads as an example, summarize and count the vehicle samples throughout the day. Among them, divide the frequency distribution of vehicles with steering behavior, observe the range of deceleration times under normal conditions. After analysis, it is found that normal steering vehicles decelerate once within one second. If there are two or more decelerations, it can be classified as a behavior-intensive section. Set the critical frequency standard of behavior as once per second. If the frequency of a certain vehicle reaches or exceeds this value, it is determined that the behavior is overly complex. Further, classify and organize the vehicle behavior information exceeding this frequency standard, and bind the position where it appears to the lane number it belongs to. Suppose a vehicle performs a left-turn action in the second lane, and its deceleration behavior exceeds the normal range and is concentrated at the end of the lane, indicating that there are interferences or obstacles affecting this position. Archive and structure this type of abnormal behavior information to form a corresponding behavior type identifier, clarify its road space position and time window, constitute the associated tag data between vehicle behavior and road section, and generate a set of interference behavior response tags.

[0068] The specific steps for obtaining the signal behavior delay section are as follows:

[0069] S201: Call the road position section identified in the interference behavior response tag group, obtain the three time data of the green light start time, the leading vehicle start time, and the trailing vehicle release time within the signal control cycle for the corresponding section, calculate the time interval between the leading vehicle response time and the trailing vehicle passing time respectively, and generate the key time interval data for passing;

[0070] Based on the section number, interference behavior timestamp, and location coordinates stored in the tag group, determine the specific road segment and direction attribute where the interference behavior occurs, retrieve the periodic release data of this section in the signal control. In the record, the time point when the green light starts, the starting time point when the first vehicle starts to move, and the ending time point when the last vehicle at the end of the queue completely exits the detection area need to be extracted for each control cycle. If in the second channel of a certain intersection, the green light starts at 8:01:10, the first vehicle starts to move at 8:01:14, and the last vehicle completely leaves the induction coil area at 8:01:28, calculate that the response time of the first vehicle is 4 seconds and the release duration of the last vehicle is 14 seconds respectively. Similar data can be obtained through automatic detection equipment or by backtracking and annotating image frames of the camera. The timestamps are recorded in milliseconds to eliminate manual errors. By batch processing the time points of each channel and each cycle, record the time interval values of the key events of the first vehicle and the last vehicle item by item, and associate each group of data with the corresponding road section and signal cycle number. Finally, summarize and output as structured data to obtain the key time interval data for passing.

[0071] S202: Based on the response time of the first vehicle and the time interval value of the passing time of the last vehicle in the key time interval data for passing, call the total number of vehicles and the vehicle position distribution within the green light cycle of the corresponding section to identify the total length of the queue and the queue sparsity level, and obtain the passing density distribution information;

[0072] Extract the number of vehicles entering and their initial arrangement position information during the green light control period under the corresponding signal cycle from the same data source. Restore the vehicle queue structure through geomagnetic coils or camera induction. For the stationary vehicles in the same lane before the green light of the current cycle, record their coordinate ranges. Mark the total length of the queue by calculating the maximum longitudinal distance from the front of the vehicle to the rear of the vehicle. At the same time, count the number of vehicles within this range and calculate the number of vehicles per unit distance to judge the sparsity of the queue. If there are 12 queuing vehicles and the total length of the queue is 96 meters, then there is 1 vehicle per 8 meters on average. According to the set standard, judge that this value is greater than the dense standard of 1 vehicle per 6 meters, which is a sparse queue. By horizontally comparing the vehicle numbers and position data sequences at different times, it is possible to further identify whether there are breakpoints or vacancies in the queue. Then, output the distribution labels of the density and sparsity of each section in matrix form, and combine and match the distribution information with the aforementioned time interval data to form the fusion analysis data of passing behavior in the two dimensions of time and space, and obtain the passing density distribution information.

[0073] S203: According to the passing density distribution information, combined with the time interval distribution of the position points in the front, middle, and rear of the queue within the key time interval for passing, calculate the passing density offset characteristic value, identify the position interval of the density abnormal section in the green light control cycle, and generate the passing density offset time period;

[0074] The formula for calculating the passing density offset characteristic value is as follows:

[0075] ;

[0076] Among them, AD represents the traffic density offset eigenvalue, n represents the number of density detection sections participating in the evaluation within the green light cycle, represents the vehicle queue length within the i-th detection section, represents the passing time length of the trailing vehicle within the i-th detection section, represents the average longitudinal response density value of the front, middle, and rear groups of vehicles in the i-th detection section, represents the delay time of the trailing vehicle in the queue section to respond to the leading vehicle in the i-th detection section, and BD represents the set reference traffic density benchmark value;

[0077] Parameter meaning and formula calculation derivation process:

[0078] During one signal control cycle, five density detection sections are obtained through road induction equipment and cameras, that is, n = 5;

[0079] In each data collection, the vehicle queue length to is obtained by multiplying the number of vehicles counted from the lane tail to the head by the headway. After actual measurement:

[0080] = 48m, = 60m, = 52m, = 44m, = 50m;

[0081] The passing time length of the trailing vehicle to is monitored by the time from when the green light comes on to when the trailing vehicle drives out of the release line. The results are:

[0082] = 10s, = 12s, = 9s, = 11s, = 10s;

[0083] The longitudinal response density values of the front, middle, and rear vehicles in each detection section to are obtained by tracking the vehicle spacing every 1 second through the camera and converting it to the average longitudinal response density value, with the unit of vehicles per second. The specific values are:

[0084] = 4.5, = 5.0, = 4.3, = 4.8, = 4.7;

[0085] Response delay time of the last vehicle to the first vehicle within the queuing section to By detecting the time difference between the start of the first vehicle and the start of the last vehicle, the collected results are as follows:

[0086] = 1.2s, = 0.8s, = 1.4s, = 0.9s, = 1.1s;

[0087] The reference traffic density benchmark value BD is determined based on the sample statistics of the green light efficient release period in the past 30 days as:

[0088] BD = 4.6;

[0089] Substitute the above values into each item of the formula and calculate them in turn:

[0090] The first part is the average absolute difference:

[0091] ;

[0092] ;

[0093] ;

[0094] ;

[0095] ;

[0096] Average after summing the first part:

[0097] ;

[0098] The second part is the root mean square of the residuals:

[0099] ;

[0100] ;

[0101] ;

[0102] ;

[0103] ;

[0104] Calculation of the root mean square of the residuals:

[0105] ;

[0106] Substitute into the formula for calculation:

[0107] AD = 0.576 + 0.4655 = 1.0415;

[0108] This result indicates that the traffic density offset eigenvalue is 1.0415. This value is greater than the offset critical judgment standard of 0.85, indicating that there is a significant offset state in the density during the current signal cycle, and the corresponding time segment needs to be marked as the traffic density offset time segment.

[0109] S204: Based on the corresponding section position in the traffic density offset time segment, evaluate the correlation between the time segment in the signal control cycle and the vehicle release behavior segment, screen out the continuous segments with offset in the vehicle release behavior, mark them as the areas where the traffic behavior is delayed, and generate the signal behavior delay section;

[0110] It is necessary to analyze each vehicle release trajectory in the signal cycle one by one, map and match the time point when the release behavior of each vehicle occurs during the passage with this time segment, and compare whether there is a trajectory segment with significant delay of a vehicle in the abnormal section. If a vehicle in a certain lane starts before the start of the density offset period, but its adjacent following vehicle is delayed for up to 6 seconds without starting, then this behavior is reflected as a release break in the trajectory. After summarizing such behaviors and classifying and screening them, identify the vehicle behavior segments that meet the characteristics of "separation of start and release", and mark their position range in the signal control cycle. By analyzing the segments, a set of traffic behaviors with a tail response time far exceeding the average response level and a significantly lagging release trajectory can be summarized, bind them with the signal cycle number and road position and output, and integrate them to form the structural identification result data to generate the signal behavior delay section.

[0111] The specific steps for obtaining the channel green light variable time period configuration table are as follows:

[0112] S301: According to the channel number indicated in the signal behavior delay section, extract the stationary duration and release completion time point of the tail vehicle of the channel at the end of the green light in two consecutive signal cycles, and combine the time difference between the release completion time point and the end time point of the green light in the current cycle to obtain the release lag information of the channel tail section;

[0113] It is necessary to retrieve the road section data calibrated by the delay tag, and further obtain the record of the status change of the trailing vehicle during the green light period of the current cycle and the previous cycle of this lane. In intelligent transportation, this type of data is jointly recorded by the video acquisition device and the induction coil at the end of the lane. The time point when the vehicle head leaves the stop line is used as the release completion mark. In actual operation, the last 5 seconds of the green light period at the end of each lane cycle is set as the late green light area. It is necessary to count whether the stationary state of the trailing vehicle continues within this interval, and the time point it takes from starting to move to completely leaving the stop line. For example, in the first cycle of a certain lane, the trailing vehicle remains stationary until the end of the green light and is not released completely. While in the second cycle, the release time of this vehicle is 2 seconds before the end of the green light, indicating that it is in a delayed release state. Compare the release completion time point with the official end time point of the green light, and record the time difference. If the difference is more than 0.5 seconds, it is classified into the lag record. The data of the trailing vehicle in each cycle is summarized on this basis, automatically recording the stationary duration and the release completion point, and forming a data structure composed of the lane number and the release difference, sorting out the release lag information at the end of the lane.

[0114] S302: Based on the release lag information at the end of the lane, judge whether the release lag value exceeds the lane release completion benchmark threshold, filter out the lane numbers with incomplete releases, and extract the waiting time of the trailing vehicles in the corresponding lanes. Sort according to the waiting duration of the incomplete lanes to generate the lane release waiting priority sequence;

[0115] During the evaluation process, it is necessary to classify the lag values of each lane, determine whether they exceed the preset lane release completion benchmark threshold. The setting of this benchmark value refers to the regional traffic flow characteristics and data samples. The lag threshold is set to 1.5 seconds for the morning peak commute period. If the trailing release lag value of a certain lane reaches 2.1 seconds, it is marked as an incomplete release lane. Based on this, filter out the lane numbers that exceed the benchmark, and call the cumulative waiting time of the corresponding trailing vehicles in the green light period of this lane to obtain the complete waiting duration of the trailing vehicle from the start of the green light to the release completion. Taking a certain lane as an example, the total waiting duration of its trailing vehicle from the start of the green light to the release completion is 17 seconds, which exceeds the upper limit of the average waiting time in the lane and needs to be considered for regulation first. Sort the lane numbers with incomplete releases in descending order according to their corresponding waiting times, establish a data pair corresponding to the lane number and the waiting duration, and mark the priority level according to the sorting number for subsequent use as the basis for the time period resource allocation order, generating the lane release waiting priority sequence.

[0116] S303: Call the sorting information in the lane release waiting priority sequence, and sequentially allocate additional green light time periods for the lanes. Adjust the length of the green light period of the lanes within the total green light duration, record the lane number and the corresponding adjusted green light duration, and generate the lane green light variable period configuration table;

[0117] According to the sorting level, additional green light time periods are allocated to the channels one by one, and the total configurable green light duration of the current signal control is re-divided and adjusted to determine the total green light variable time budget, which is set to 20 seconds. The time is allocated to the channels with higher priority in a larger proportion, and then reduced in sequence. If the first ranked channel is numbered A01, an additional 6 seconds of green light can be configured for it, and the second and third channels are allocated 5 seconds and 4 seconds respectively. The remaining channels are allocated according to the remaining budget. Each allocation records the channel number, the sum of the original green light duration and the newly configured time, and updates the green light time of the channel in the next cycle. In the actual configuration table, each record includes five data fields: channel number, current signal cycle number, original duration, newly added duration and merged green light duration. Structured output is automatically generated, and the configuration item interface parameters of the corresponding section are established to ensure matching with the traffic control terminal, and generate a channel green light variable time period configuration table.

[0118] The specific steps for obtaining the signal cycle start instruction are:

[0119] S401: calling the channel start configuration time value recorded in the channel green light variable period configuration table, detecting the longitudinal distance distribution of vehicles in the corresponding channel queue area within the real-time signal cycle, identifying the continuous acceleration start time points of three adjacent vehicles in the queue section, and generating the queue acceleration start time;

[0120] According to each record in the configuration table, the channel number and green light start time information are extracted, and the specific starting point of the green light section of the channel in the current cycle is located. The position coordinates and identification numbers of the stationary or moving vehicles in the channel queue area are obtained through continuous frame cameras or geomagnetic coils deployed at the entrance and middle area of ​​the lane, and the center point position of each vehicle is counted with the longitudinal coordinate as the main axis to form a longitudinal distance distribution map. In the distribution map, samples are taken at intervals of 1 second, and the longitudinal coordinates of the vehicles are compared in sequence to identify whether the longitudinal displacement of the vehicle in two consecutive frames is greater than the set minimum acceleration threshold. If the displacement of vehicles B, C, and D is continuously greater than 1 meter, they are marked as continuous acceleration states, and the time points at which each of them first appears to meet this condition are recorded, that is, the acceleration start time points. It is set that vehicle B accelerates in the 2nd second after the green light starts, C in 2.2 seconds, and D in 2.4 seconds. All three are in a sequential acceleration state. The timestamp is recorded as an acceleration group event, and its channel number, vehicle number, and acceleration start time are constructed together as a data set structure to generate the queue acceleration start time.

[0121] S402: judging whether the acceleration time is earlier than the signal preparation switching time point in the variable period of the green light of the channel according to the time points of the acceleration action of the three adjacent vehicles in the queuing acceleration start time, and if the judgment condition is met, marking it as a synchronous start state, binding the channel number and state information, and generating a signal cycle start instruction;

[0122] The formula for determining whether the acceleration time is earlier than the signal preparation switching time point within the variable green light period of the channel is as follows:

[0123] ;

[0124] Among them, AT represents the average starting timing difference, 、 、 respectively represent the acceleration starting times of the first, second, and third vehicles in this group, represents the speed at the sampling moment before the acceleration action of the a-th vehicle, OV represents the passing reference speed set for the road section, k is the weighting coefficient, and m is the number of vehicles;

[0125] Parameter meaning and formula calculation derivation process:

[0126] 、 、 : Through a non-contact speedometer or a drum test bench with high-frequency sampling (≥10Hz), record the speed-time curve of the vehicle during the starting acceleration process. When the vehicle speed rises from 0 km / h to 1 km / h for the first time, record this moment as the acceleration starting time. Set the acceleration starting times of the three vehicles as follows:

[0127] The first vehicle (vehicle 1): = 0.8 s;

[0128] The second vehicle (vehicle 2): = 1.1 s;

[0129] The third vehicle (vehicle 3): = 1.4 s;

[0130] 、 、 : Through an on-vehicle wheel speed sensor or GPS, obtain the speed value at the last sampling moment before the vehicle accelerates. Set the speeds of the three vehicles as follows:

[0131] Vehicle 1: = 0.5 m / s;

[0132] Vehicle 2: = 0.6 m / s;

[0133] Vehicle 3: = 0.4 m / s;

[0134] OV: The designed passing reference speed for the urban arterial road is 60 km / h, that is: OV = 60 km / h = 16.67 m / s;

[0135] k: The weighting coefficient, which is used to correct the influence of speed difference on the starting time difference. It is set based on the sensitivity analysis of the speed difference to the time difference and is set as: k = 0.05;

[0136] Formula calculation process:

[0137] Calculate the time difference:

[0138] ;

[0139] ;

[0140] Calculate the speed difference term:

[0141] ;

[0142] ;

[0143] Calculate the square root term:

[0144] ;

[0145] ;

[0146] Substitute into the formula for calculation:

[0147] ;

[0148] This result shows that the average starting time sequence difference of the three vehicles is 1.348 seconds. If this value is less than the set synchronous start threshold (for example, 2 seconds), it can be determined that the current queue is in a synchronous start state, and a signal cycle start instruction is generated accordingly.

[0149] The specific steps to obtain the regulation result of the green light duration at the end section are as follows:

[0150] S501: Call the signal cycle number marked in the signal cycle start instruction, filter the trajectory data of the tail vehicle in the latter part of the green light under the corresponding cycle, extract the continuous speed sequence and timestamp data from the starting moment of the end section to the moment when the vehicle exits the release line, and generate the end section passing trajectory sequence;

[0151] First, the control cycle information corresponding to the current analysis object needs to be matched from the signal regulation data table, and then the passing trajectory records of the trailing vehicle in this cycle are screened from the vehicle trajectory database. In actual operation, through the geomagnetic sensors or camera recognition arranged in the detection area at the tail section, the speed changes and time information of the last three vehicles in the last stage of the green light are continuously recorded. The sampling frames are aligned based on the starting time of the green light, and the starting time of the tail section is defined as the stage after the fifth second before the end of the green light countdown. During this time period, the motion state of each vehicle is sampled at intervals of 0.5 seconds, the speed change values and corresponding timestamps are extracted, and a complete time series is formed. For example, a trailing vehicle samples its speed every 0.5 seconds starting from the 7th second of the green light, and a total of ten sets of speed data are obtained, constructing a complete passing trajectory time series within the tail section. This time series is archived with the vehicle number as the primary key, marked with the channel number, signal cycle number, and start and end times of the frame, and the tail section passing trajectory sequence is output.

[0152] S502: Based on the speed sequence in the tail section passing trajectory sequence, analyze the speed change trend of the trailing vehicle during the green light duration in the tail section, extract the speed increase value at the end of the sequence, and compare it with the corresponding duration of the tail section. If the speed continues to rise and has not entered the stable range, calculate the additional time required for the vehicle to reach the predetermined release point, and generate the remaining release time in the tail section;

[0153] It is necessary to conduct a trend analysis on the motion characteristics of the trailing vehicle during the last stage of the green light to determine whether the vehicle is in a continuous acceleration state and has not entered the stable operation range. The specific analysis steps are as follows: Take three consecutive sets of data from the end of the speed sequence and calculate the increase value and compare it with the previous speed increase amplitude. If the difference maintains a positive growth trend and the change does not tend to converge, it is determined that the vehicle has not passed stably. If the speed of the vehicle increases from 5.6 m / s to 6.1 m / s and then to 6.7 m / s in the last three frames, and the continuous increase amplitude exceeds 0.5 m / s each time, it is judged that the vehicle is still in the acceleration state. Then, compare it with the remaining time in the tail section. If the remaining time in the current last stage of the green light is only 1.5 seconds, and the vehicle still needs 2.3 seconds to reach the release point, then 0.8 seconds of additional time needs to be added. Record this time value and archive it together with the channel number and vehicle number. Each set of data generated includes the acceleration trend state, the speed increase at the end of the section, the additional time, and the signal cycle number, forming the remaining release time in the tail section.

[0154] S503: Call the additional time value required for the channel in the remaining release time in the tail section, update the real-time green light control cycle in the tail section, correct the original configured end time point of the green light in the tail section, and reset the control interval of the tail section of the channel to obtain the continuous regulation result of the green light in the tail section;

[0155] The original green light control configuration for the rear section will be dynamically adjusted in the current signal cycle. The end time point of each channel's rear section in the current cycle is read, and then the end time is updated one by one according to the supplementary time recorded in the remaining release time of the rear section. The set end time of the current rear section of channel number B03 is set to the 14th second of the green light. After determining that 1.2 seconds need to be supplemented, its end point is extended to 15.2 seconds. This operation needs to consider whether the total cycle time is sufficient and complete the duration allocation on the basis of not affecting the channel control rhythm. Each adjustment records the timestamps before and after the modification, and synchronously updates the controller output data. The configuration results are structured and sorted according to four data items: channel number, cycle number, original end time, and new end time, and a control parameter configuration file in a standardized format is output for real-time update and scheduling by the signal control terminal, and the continuous regulation result of the rear section green light is output.

[0156] Please refer to Figure 2 , an intelligent traffic signal parameter configuration system, which includes:

[0157] The behavior data recognition module detects the vehicle passing behaviors of the lanes at the signal-controlled intersection, collects the lateral offset positions and longitudinal delay time series of the vehicles, counts the deceleration frequencies of the vehicles, and generates an interference behavior response tag group;

[0158] The interference behavior analysis module, based on the interference behavior response tag group, identifies the key behavior actions that cause traffic interruption, binds the key behavior actions to the locations, and generates signal behavior delay sections;

[0159] The signal delay calculation module calls the signal behavior delay sections, obtains the green light start time, the starting time of the first vehicle, and the release time of the last vehicle within the sections, calculates the time intervals, and generates a channel green light variable time period configuration table;

[0160] The green light time dynamic configuration module, based on the channel green light variable time period configuration table, analyzes the completion situation of the channel release, calculates the waiting duration of the uncompleted channels, reallocates the green light time periods as needed, and generates a signal cycle start instruction;

[0161] The signal start module calls the signal cycle start instruction, detects and evaluates the longitudinal distances and continuous acceleration points of the queued vehicles, adjusts the signal start moment according to the vehicle speed change trend and the rear section duration, and obtains the continuous regulation result of the rear section green light.

[0162] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for configuring intelligent traffic signal light parameters, characterized in that: The following steps are involved: S1: Obtain vehicle traffic data of lanes at signal-controlled intersections, extract the lateral offset position and longitudinal delay sequence before and after the turning action, make a judgment based on the vehicle deceleration frequency corresponding to the behavior, and generate an interference behavior response label group; S2: calling the marked position in the interference behavior response tag group, obtaining the green light start time, the first vehicle start time and the last vehicle release time in the section, identifying the behavior and traffic deviation fragments in the signal control cycle, and generating a signal behavior delay section; S3: According to the channel indicated in the signal behavior delay section, the waiting time of the rear vehicle at the end of the green light and the release completion time point within two cycles of the channel are extracted, the time difference between the end point of the real-time green light is calculated, and it is determined whether the channel release is not completed. The waiting time corresponding to the unfinished channel is selected as the priority sorting basis, and the channel green light variable time period configuration table is generated; S4: Call the starting configuration value in the channel green light variable period configuration table, detect the longitudinal distance of the vehicles in the queue in the real-time cycle, and mark the continuous acceleration time points of the three adjacent vehicles. If the acceleration time is lower than the signal preparation switching point, record the synchronous start status and generate a signal cycle start instruction.

2. The intelligent traffic signal light parameter configuration method according to claim 1, characterized in that: The interference behavior response label group includes the traffic interruption identification number, the interference behavior classification code, and the lane space mapping position; the signal behavior delay section includes the signal start response delay length, the vehicle release rhythm lag length, and the behavior trigger and traffic rhythm offset length; the channel green light variable time period configuration table includes the channel sequence sorting weight, the green light time period adjustment amplitude, and the scheduling control effective time; the signal cycle start instruction includes the cycle switching trigger state, the synchronous start trigger time, and the vehicle distance compression recognition state.

3. The intelligent traffic signal light parameter configuration method according to claim 1, characterized in that: The steps for obtaining the interference behavior response label group are specifically as follows: S101: Acquire vehicle traffic data of lanes at signal-controlled intersections, extract the lateral coordinate change and longitudinal position change time sequence of each vehicle before and after turning, calculate the lateral offset distance and longitudinal action delay period based on the time length of each lateral coordinate difference and longitudinal sequence, and generate an offset and delay time sequence parameter set; S102: according to the lateral offset distance and longitudinal action delay period of the vehicle in the offset and delay timing parameter set, combined with the speed change information of the vehicle at consecutive moments before and after the turning action, counting the number of decelerations of each vehicle in the offset and delay stage, identifying the synchronous distribution relationship between the number of decelerations and the delay period, and generating the deceleration frequency of the vehicle in the turning stage; S103: Based on the deceleration frequency of the vehicle in the turning phase, determine the vehicle behavior whose deceleration frequency exceeds the preset critical frequency of the turning behavior, bind the corresponding behavior action with the road section position, and generate an interference behavior response label group.

4. The intelligent traffic signal light parameter configuration method according to claim 3, characterized in that: The steps for obtaining the signal behavior delay section are specifically as follows: S201: calling the road position section identified in the interference behavior response tag group, obtaining three time data of the green light start time, the first vehicle start time and the last vehicle release time of the corresponding section within the signal control cycle, respectively calculating the first vehicle response time and the last vehicle passing time interval, and generating the key time interval data for passing; S202: Based on the response time of the first vehicle and the time interval of the last vehicle in the critical time interval data, the total number of vehicles and the vehicle position distribution in the green light cycle of the corresponding section are called, the total length of the queue and the sparseness level of the queue are identified, and the traffic density distribution information is obtained; S203: Calculate the traffic density offset characteristic value based on the traffic density distribution information and the time interval distribution of the front, middle and rear position points of the queue in the critical traffic time interval, identify the position interval of the density abnormal section in the green light control cycle, and generate the traffic density offset time period; The formula for calculating the traffic density offset characteristic value is as follows: ； Among them, AD represents the characteristic value of traffic density deviation, n represents the number of density detection segments involved in the evaluation within the green light cycle, represents the length of the vehicle queue in the i-th detection segment, represents the length of time the tail vehicle passes through in the i-th detection segment, Represents the average longitudinal response density value of the front, middle and rear three groups of vehicles in the i-th detection segment, represents the delay time of the tail vehicle in the queue section in the i-th detection section responding to the head vehicle, and BD represents the set reference traffic density benchmark value; S204: Based on the corresponding segment position in the traffic density offset time period, the correlation between the time segment and the vehicle release behavior segment within the signal control cycle is evaluated, and the continuous segments with offset in the vehicle release behavior are screened and marked as the traffic behavior delay occurrence area, and the signal behavior delay segment is generated.

5. The intelligent traffic signal light parameter configuration method according to claim 4, characterized in that: The steps for obtaining the channel green light variable time period configuration table are specifically as follows: S301: According to the channel number indicated in the signal behavior delay section, extract the stationary duration and release completion time point of the rear vehicle at the end of the green light in two consecutive signal cycles of the channel, and combine the time difference between the release completion time point and the end time point of the green light in the current cycle to obtain the release hysteresis information of the end of the channel; S302: Based on the release hysteresis information of the tail section of the channel, determine whether the release hysteresis value exceeds the channel release completion reference threshold, filter the channel numbers of the uncompleted releases, extract the waiting time of the tail vehicles in the corresponding channels, sort the uncompleted channels according to the waiting time, and generate a channel release waiting priority sequence; S303: Call the sorting information in the channel release waiting priority sequence, configure additional green light time periods for the channels in sequence, adjust the channel green light time period length within the total green light duration, record the channel number and the corresponding adjusted green light duration, and generate a channel green light variable time period configuration table.

6. The intelligent traffic signal light parameter configuration method according to claim 5, characterized in that: The steps for obtaining the signal cycle start instruction are specifically as follows: S401: calling the channel start configuration time value recorded in the channel green light variable period configuration table, detecting the longitudinal distance distribution of vehicles in the corresponding channel queue area within the real-time signal cycle, identifying the continuous acceleration start time points of three adjacent vehicles in the queue section, and generating the queue acceleration start time; S402: According to the time points of the acceleration actions of the three adjacent vehicles in the queue acceleration start time, determine whether the acceleration time is earlier than the signal preparation switching time point within the variable period of the channel green light. If the judgment condition is met, it is marked as a synchronous start state, and the channel number and status information are bound to generate a signal cycle start instruction.

7. The intelligent traffic signal light parameter configuration method according to claim 6, characterized in that: The formula for judging whether the acceleration time is earlier than the signal preparation switching time point within the variable period of the channel green light is as follows: ； Among them, AT represents the average starting timing difference, , , Respectively represent the acceleration start time of the first, second, and third vehicles in the group. It represents the speed of the a-th vehicle at the sampling time before the acceleration action, OV represents the reference speed set for the road section, k is the weighting coefficient, and m is the number of vehicles.

8. The intelligent traffic signal light parameter configuration method according to claim 1, characterized in that: The method further comprises step S5: S5: calling the marking cycle of the signal cycle start instruction, screening the passage trajectory of the rear green light of the rear vehicle, comparing the vehicle speed change trend with the tail segment duration, and if the speed keeps rising but does not reach stability, calculating the time required to reach the predetermined release point and updating the tail segment control cycle to obtain the tail segment green light continuous control result; The result of the continuous control of the tail green light includes the duration of the remaining release path, the predicted time period for the completion of the tail vehicle release, and the recommended time window for signal maintenance.

9. The intelligent traffic signal light parameter configuration method according to claim 8, characterized in that: The steps for obtaining the continuous control result of the tail green light are specifically as follows: S501: calling the signal cycle number marked in the signal cycle start instruction, filtering the trajectory data of the rear vehicle in the rear section of the green light under the corresponding cycle, extracting the continuous speed sequence and timestamp data from the start time of the rear section to the vehicle driving out of the release line, and generating the rear section passing trajectory sequence; S502: Based on the speed sequence in the tail passage trajectory sequence, the speed change trend of the tail vehicle within the green light duration of the tail segment is analyzed, the speed increase value of the tail segment of the sequence is extracted, and compared with the duration corresponding to the tail segment. If the speed continues to increase and does not enter the stable interval, the time required for the vehicle to reach the predetermined release point is calculated to generate the remaining release time of the tail segment; S503: Call the required supplementary time value of the channel in the remaining release time interval of the tail segment, update the real-time tail segment green light control cycle, correct the originally configured tail segment green light end time point, reset the control interval of the channel tail segment, and obtain the tail segment green light continuous control result.

10. An intelligent traffic light parameter configuration system, characterized in that: The system is used to implement the intelligent traffic signal light parameter configuration method according to any one of claims 1 to 9, and the system includes: The behavior data recognition module detects the vehicle traffic behavior in the lane of the signal-controlled intersection, collects the lateral offset position and longitudinal delay time sequence of the vehicle, counts the vehicle's deceleration frequency, and generates an interference behavior response label group; The interference behavior analysis module identifies the key behavior actions that cause traffic interruption based on the interference behavior response tag group, binds the key behavior actions to locations, and generates a signal behavior delay section; The signal delay calculation module calls the signal behavior delay section, obtains the green light start time, the first vehicle start time and the last vehicle release time in the section, calculates the time interval, and generates a channel green light variable time period configuration table; The green light time dynamic configuration module analyzes the completion of channel release based on the channel green light variable time period configuration table, calculates the waiting time of the unfinished channel, reallocates the green light time period as needed, and generates a signal cycle start instruction; The signal start module calls the signal cycle start instruction, detects and evaluates the longitudinal distance and continuous acceleration points of the vehicles in the queue, adjusts the signal start time according to the vehicle speed change trend and the tail segment duration, and obtains the tail segment green light continuous control result.

Images