A large model data collaborative management system and method for traffic

Abstract

The application discloses a large model data cooperative management system and method for traffic, and relates to the technical field of traffic management, which comprises the following steps: dividing a day into multiple time windows; obtaining a target intersection to be adjusted in timing, dividing a target intersection area and each lane area connected with the target intersection, and establishing a turning proportion set of the lane at the target intersection; establishing a passing function of the number of passing vehicles changing with time; determining a current timing green light phase to be adjusted and an analysis lane to be analyzed, calculating a target distance and a target cross section, and calculating the number of unit vehicles passing the target cross section; calculating reliable values of the timing green light phase to be adjusted when different timing durations are adopted, and automatically recommending a timing duration of the timing green light phase. Through the analysis of the change trend of intersection vehicles and the turning behavior of the vehicles, the timing of the signal lamp is intelligently adjusted, the problem that the timing strategy does not match the actual passing demand is solved, the overall passing efficiency of the intersection is effectively improved, and the passing experience and satisfaction of the drivers are improved.

Description

Technical Field

[0001] This invention relates to the field of traffic management technology, specifically a large-scale model data collaborative management system and method for traffic. Background Technology

[0002] The current urban traffic signal timing mechanism has significant shortcomings. Most intersections operate according to a preset fixed timing scheme, which cannot be dynamically adjusted according to traffic flow fluctuations at different times and the actual traffic volume and queuing status of each lane. This easily leads to a mismatch between the timing strategy and actual traffic demand. For example, at certain times, some lanes have high traffic volume, but the green light duration is insufficient to disperse all vehicles within a single signal cycle, resulting in long waiting times. At other times, the lanes have sparse traffic, but still occupy the fixed green light duration, resulting in wasted green light and lane resources. Therefore, the existing method is unable to achieve intelligent adaptive adjustment of traffic lights, causing an imbalance in the allocation of lane traffic resources. This not only reduces the overall traffic efficiency of the intersection and exacerbates traffic congestion, but also significantly affects the driving experience and satisfaction of drivers. Summary of the Invention

[0003] The purpose of this invention is to provide a large-scale model data collaborative management system and method for transportation, in order to solve the problems raised in the prior art.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A collaborative management method for large-scale model data in transportation includes the following steps: A fixed-duration time interval is used to divide a day into multiple time windows; Match the current time to the time window, obtain the target intersection to be timed, and retrieve the historical monitoring segments of the target intersection within the time window; divide the target intersection area and the lane areas connected to the target intersection in the monitoring segments, and label the vehicles in the monitoring segments; analyze the turning behavior of vehicles at the target intersection, and establish a set of turning ratios for each lane at the target intersection. The target intersection refers to the intersection studied in this scheme that needs to be re-timed. The lane refers to the motor vehicle lane connected to the target intersection. In this scheme, the lane contains several lanes that can travel in the same direction. Some lanes can allow vehicles to make U-turns or / and turn left, while some lanes can allow vehicles to go straight or / and turn right. Some lanes can allow U-turns, left turns, and also go straight and turn right. However, since the probability of a vehicle turning left, going straight, turning right, and making a U-turn after passing the target intersection is a relatively fixed empirical probability value, which can be obtained based on the statistics of historical traffic flow data, this scheme needs to obtain the set of turning ratios of the lanes at the target intersection to make the following calculations reliable. Obtain the number of vehicles passing through the target intersection in the historical monitoring video and the passing duration, and establish a traffic function describing the change of the number of passing vehicles over time; Determine the current green light phase to be timed, take the passable lanes controlled by the green light phase to be timed as lanes to be analyzed, calculate the target distance corresponding to the lanes to be analyzed; determine the target section according to the target distance, and calculate the number of unit vehicles passing through the target section based on vehicles traveling on the lanes to be analyzed; Calculate the reliability value when the green light phase to be timed has different timing durations according to the turning ratio set, the traffic function and the number of unit vehicles, and automatically recommend the timing duration of the green light phase based on the reliability value.

[0005] Preferably, establishing a turning ratio set for each lane at the target intersection includes: Divide the target intersection area and each connected lane area in the monitoring segment, mark vehicle C in the monitoring segment, which travels from lane L, passes through the target intersection and heads for another lane, use a target detection algorithm to divide the area where vehicle C is located, detect the first direction when vehicle C just touches the target intersection and the second direction when it exits the target intersection, and determine the turning behavior of vehicle C according to the first direction and the second direction; Determine the probability of each turning behavior according to the turning behavior of each vehicle that travels from lane L, passes through the target intersection and heads for other lanes, and establish the turning ratio set of lane L at the target intersection based on this.

[0006] Preferably, establishing the traffic function describing the change of the number of passing vehicles over time includes: Extract historical monitoring segments of the target intersection, obtain the green light cycle allowing vehicles on lane L to pass and the red light cycle prohibiting vehicles from passing, obtain the to-be-driven vehicles waiting on lane L during the red light cycle, record the number as M, take the to-be-driven vehicles that leave lane L during the next adjacent green light cycle after the red light cycle as passing vehicles, record the number of passing vehicles as m, where 0<m≤M, and record the duration from the moment when the green light starts to the moment when the last passing vehicle leaves lane L as the passing duration; Determine a plurality of numbers of passing vehicles and passing durations according to the corresponding plurality of red light cycles and green light cycles in the monitoring segments, perform linear function fitting, and establish the traffic function of the number of passing vehicles varying with the passing duration.

[0007] The standard form of the linear function is y=kx+b, where y is the dependent variable, x is the independent variable, k is the slope, and b is the intercept. In this solution, the dependent variable is the number of passing vehicles, and the independent variable is the passing duration. Under normal circumstances, since the number of passing vehicles increases with the increase of the passing duration, the traffic function can characterize the linear relationship between the number of passable vehicles and the vehicle passing duration at the target intersection.

[0008] Preferably, the calculation of the number of vehicles per unit passing through the target section includes: Retrieve the lane to be analyzed and the passable turning behavior corresponding to the green light phase when the lane to be analyzed is to be matched, and calculate the target distance corresponding to the lane to be analyzed based on the historical monitoring video of the lane to be analyzed. Divide the lane to be analyzed into a target section that is the same distance from the target intersection. Deploy monitoring equipment at the target section and retrieve historical monitoring segments that are consistent with the current time window. Based on the number of vehicles Q passing through the target section in the monitoring segment and the duration D of the monitoring segment, obtain the number of vehicles per unit passing through the target section N0 = Q / D.

[0009] Preferably, the calculation of the target distance corresponding to the lane to be analyzed includes: Obtain the allowable timing range of the green light phase to be matched [D] min D max ], where D min D is the preset minimum timing duration. max The preset maximum timing duration; Calculate the average speed of vehicles traveling in the lane to be analyzed, multiply it by the shortest timing duration, and obtain the target distance corresponding to the lane to be analyzed.

[0010] Preferably, the green light phase timing duration is automatically recommended based on reliability values, including: Substituting the timing duration t of the green light phase to be scheduled into the traffic function, we obtain the first number of vehicles M. t ; Using the end time of the immediately preceding red light phase of the green light phase to be scheduled as the statistical time, based on the surveillance video, count the number of vehicles N1 between the target section and the target intersection. Based on the number of vehicles per unit N0, calculate the expected number of vehicles N=N0(tD) when the green light phase to be scheduled has a duration of t. min )+N1,D min To determine the minimum preset timing duration, obtain the steering behavior probability of the lane to be analyzed during the green light phase of the desired timing from the set of steering ratios. Multiply the expected number of vehicles N by the corresponding steering behavior probability to obtain the second number of vehicles N. t And calculate the target value X=N t / M t ; The reliable value for calculating the timing duration t is Calculate the reliability values ​​corresponding to multiple timing durations, and take the timing duration corresponding to the maximum reliability value as the recommended timing duration for the green light phase to be timed.

[0011] It should be noted that the formula Z is a function of X > 0, where Z ranges from 0 to 1. Specifically, when X is between 0 and 1, Z increases with increasing X; when X > 1, Z decreases with increasing X; and Z has a maximum value of 1 when X = 1. To achieve a reasonable duration setting, this application aims to maximize X = N. t / M t It's close to a 1 match because X=N t / M t The value of reliability Z is closer to 1. The larger the reliability Z value, the more reasonable the timing duration is. Therefore, it can be used as the recommended timing duration for the green light phase waiting to be matched.

[0012] A large-scale model data collaborative management system for traffic includes a steering ratio set establishment module, a traffic function establishment module, a unit vehicle quantity calculation module, and a timing duration recommendation module; Steering ratio set establishment module: used to divide a day into multiple time windows using a fixed duration time interval division method; Match the current time to the time window, obtain the target intersection to be timed, and retrieve the historical monitoring segments of the target intersection within the time window; divide the target intersection area and the lane areas connected to the target intersection in the monitoring segments, and label the vehicles in the monitoring segments; analyze the turning behavior of vehicles at the target intersection, and establish a set of turning ratios for each lane at the target intersection. The passage function establishment module is used to obtain the number of vehicles passing through the target intersection and the passage time in historical surveillance videos, and to establish a passage function that changes the number of passing vehicles over time. Unit vehicle quantity calculation module: used to determine the current green light phase to be scheduled, take the passable lane controlled by the green light phase to be scheduled as the lane to be analyzed, calculate the target distance corresponding to the lane to be analyzed; determine the target section based on the target distance, and calculate the unit vehicle quantity passing through the target section based on the vehicles traveling on the lane to be analyzed. Timing Recommendation Module: This module calculates reliable values ​​for different timing durations of the green light phase based on the steering ratio set, traffic function, and number of vehicles per unit, and automatically recommends green light phase timing durations based on these reliable values.

[0013] Preferably, the steering ratio set establishment module includes a steering behavior determination unit and a steering ratio set establishment unit; Steering behavior determination unit: used to divide the target intersection area and the connected lane areas in the monitoring segment, label the vehicles in the monitoring segment, use the target detection algorithm to divide the area where the vehicle is located, detect the first direction when the vehicle just touches the target intersection, and the second direction when it leaves the target intersection, and determine the vehicle's steering behavior based on the first and second directions. Steering Proportion Set Establishment Unit: Used to determine the probability of each steering behavior based on the steering behavior of each vehicle coming from the lane and passing through the target intersection to other lanes, and to establish the steering proportion set of the lane at the target intersection.

[0014] Preferably, the unit vehicle quantity calculation module includes a target distance determination unit and a unit vehicle quantity calculation unit; Target distance determination unit: used to retrieve the lane to be analyzed and the passable turning behavior corresponding to the green light phase to be matched, and calculate the target distance corresponding to the lane to be analyzed based on the historical monitoring video of the lane to be analyzed; Unit vehicle quantity calculation unit: used to divide the target section within the lane to be analyzed, with a length equal to the target distance from the target intersection, deploy monitoring equipment at the target section, retrieve monitoring segments from the same historical time period that are consistent with the current time window, and obtain the unit vehicle quantity passing through the target section based on the number of vehicles passing through the target section in the monitoring segment and the duration of the monitoring segment.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention provides a large-scale model data collaborative management system and method for traffic, including: dividing a day into multiple time windows; acquiring the target intersection to be timed, dividing the target intersection area and the lane areas connected to the target intersection, and establishing a set of lane turning ratios at the target intersection; establishing a traffic function for the number of passing vehicles changing over time; determining the current green light phase to be timed and the lane to be analyzed, calculating the target distance and target cross-section, and calculating the number of vehicles per unit passing through the target cross-section; calculating the reliable value of the green light phase to be timed for different timing durations, and automatically recommending the green light phase timing duration. This invention intelligently adjusts traffic light timing by analyzing the changing trends of vehicles at intersections and their turning behavior, solving the problem of mismatch between timing strategies and actual traffic demand, effectively improving the overall traffic efficiency of intersections, and enhancing the driving experience and satisfaction of drivers. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a collaborative management method for large-scale model data in transportation according to the present invention. Detailed Implementation

[0018] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure.

[0019] Example: Figure 1 As shown, this invention provides a technical solution for collaborative management of large-scale model data in transportation, comprising the following steps: A fixed-duration time interval is used to divide the day into multiple time windows. Since the vehicle density of each lane varies greatly at different times, it is necessary to conduct traffic flow characteristic analysis for each time window to realize the time-sharing dynamic adjustment of traffic light timing and formulate a differentiated reliable green light timing strategy. In this embodiment, the fixed-duration time interval is 1 hour, that is, the day is divided into 24 time windows according to 00:00-01:00, 01:00-02:00, ..., 23:00-24:00. It should be noted that traffic flow at urban intersections exhibits a clear intraday cyclical characteristic, forming differentiated traffic periods such as morning peak, evening peak, daytime off-peak, and nighttime low-peak. The purpose of dividing time windows in this scheme is primarily to calculate the set of turning proportions of vehicles passing through the target intersection within different time windows. Therefore, if the granularity of the time window is too large, meaning the duration of a single time window is too long, it will merge traffic flow samples with completely different driving path preferences, such as morning peak, off-peak, evening peak, and low-peak, for statistical analysis. Since the turning patterns of vehicles differ significantly between morning peak and off-peak scenarios, direct superposition will greatly amplify the calculation error of the turning proportion set, leading to an imbalance in the accuracy of subsequent traffic light timing calculations. On the other hand, if the granularity of the time window is too small, meaning the duration of a single time window is too short, it will lead to excessive computational pressure on the computer system, causing system task backlog and increased runtime latency. Therefore, in order to balance statistical accuracy and system computational load, this embodiment sets the granularity of the time window to 1 hour. Of course, 1 hour is only a preferred implementation method of this scheme, and the duration of the time window can be flexibly adjusted in actual applications. This application does not impose a unique limitation on this.

[0020] Match the current time to its corresponding time window, obtain the target intersection to be timed, and retrieve historical monitoring segments of the target intersection within that time window; divide the target intersection area and the lane areas connected to the target intersection within the monitoring segments, and label the vehicles in the monitoring segments; analyze the turning behavior of vehicles at the target intersection, and establish a set of turning ratios for each lane at the target intersection, specifically: It should be noted that the target intersection refers to the intersection studied in this scheme that requires rescheduling, and the lane refers to the motor vehicle lane connected to the target intersection. In this scheme, the lane contains several lanes that can travel in the same direction. Some lanes can allow vehicles to make U-turns or / and turn left, while others can allow vehicles to go straight or / and turn right. Some lanes can allow U-turns, left turns, and also go straight and turn right. Since, under normal circumstances, the probability of a vehicle turning left, going straight, turning right, or making a U-turn after passing the target intersection is a relatively fixed empirical probability value, which can be obtained from the statistics of a large amount of historical traffic flow data, this scheme needs to obtain the set of turning ratios of the lanes at the target intersection to ensure the reliability of the following calculations.

[0021] The monitoring segment is divided into the target intersection area and the adjacent lane areas. The YOLO object detection algorithm is used to detect vehicles within the monitoring segment, outputting the bounding box of each vehicle and labeling vehicle C. Vehicle C originates from lane L, passes the target intersection, and moves to another lane. The specific implementation of the YOLO object detection algorithm follows existing steps and will not be elaborated here. Since there are many vehicles at road intersections in real traffic scenarios, such as vehicles A, B, and D in addition to vehicle C, this embodiment outputs the bounding boxes of all vehicles within each frame using YOLO, and then integrates them with ByteTrack. A multi-target tracking algorithm locks onto vehicle C and continuously outputs its bounding box in each frame of the video sequence, thus obtaining the position of vehicle C in each frame of the monitored segment. It then detects the first direction when vehicle C first approaches the target intersection and the second direction when it exits the intersection. Based on the first and second directions, it determines the turning behavior of vehicle C. Common sense dictates that the first and second directions characterize the actual turning of the vehicle at the intersection. In this embodiment, the direction pointed to by the first direction is taken as the positive reference direction, and the angle range of the second direction is determined clockwise. If the clockwise angle is between 60° and 120°, the vehicle's turning behavior is determined to be a left turn; if the clockwise angle is between 150° and 210°, the vehicle's turning behavior is determined to be going straight; if the clockwise angle is between 240° and 300°, the vehicle's turning behavior is determined to be a right turn; if the clockwise angle is between 0° and 30° or between 330° and 360°, the vehicle's turning behavior is determined to be a U-turn.

[0022] Based on the turning behavior of each vehicle coming from lane L and passing through the target intersection to other lanes, determine the probability of each turning behavior, and establish the turning ratio set of lane L at the target intersection. Here, the turning ratio set is [g1, g2, g3, g4], where g1, g2, g3, and g4 are the probabilities of left turn, straight, right turn, and U-turn turning behaviors, respectively, and satisfy g1+g2+g3+g4=1.

[0023] Obtain the number of vehicles passing through a target intersection and their passing duration from historical surveillance videos, and establish a traffic function describing the change in the number of passing vehicles over time, specifically as follows: Extract historical surveillance clips of the target intersection, obtain the green light cycle during which vehicles are allowed to pass on lane L and the red light cycle during which vehicles are not allowed to pass, obtain the number of to-be-driven vehicles waiting on lane L during the red light cycle, record the number as M, take the to-be-driven vehicles that leave lane L during the next green light cycle immediately adjacent to the red light cycle as passing vehicles, record the number of passing vehicles as m, where 0<m≤M, and record the duration from the moment the green light starts to the moment the last passing vehicle leaves lane L as the passing duration; According to a plurality of corresponding red light cycles and green light cycles in the surveillance clips, determine a plurality of numbers of passing vehicles and passing durations, perform linear function fitting, and establish the traffic function representing the change of the number of passing vehicles with the passing duration.

[0024] The standard form of the linear function is y=kx+b, where y is the dependent variable, x is the independent variable, k is the slope, and b is the intercept. In this solution, the dependent variable is the number of passing vehicles, and the independent variable is the passing duration. Under normal conditions, since the number of passing vehicles increases as the passing duration increases, the traffic function can represent the linear relationship between the number of passable vehicles and the vehicle passing duration at the target intersection.

[0025] Determine the current green light phase to be timed, take the passable lane controlled by the green light phase to be timed as the lane to be analyzed, calculate the target distance corresponding to the lane to be analyzed; determine the target section according to the target distance, and calculate the number of unit vehicles passing through the target section based on the vehicles traveling on the lane to be analyzed, specifically as follows: Retrieve the lane to be analyzed and the allowed passing turning behaviors corresponding to the green light phase to be timed, and calculate the target distance corresponding to the lane to be analyzed according to the historical surveillance video of the lane to be analyzed, comprising: Obtain the allowable timing duration range [D min ,D max , wherein D min is a preset minimum timing duration, D max is a preset maximum timing duration; the specific allowable timing duration range [D min ,D max is determined according to actual conditions. In this embodiment, D min is 20 seconds, and D max is 100 seconds.

[0026] Calculate the average speed of vehicles traveling on the lane to be analyzed, multiply the average speed by the minimum timing duration to obtain the target distance corresponding to the lane to be analyzed.

[0027] Here, the average speed is V, and the shortest timing duration is D. min According to V multiplied by D min The significance of determining the target distance and then determining the target cross-section based on the target distance is: since the shortest timing duration of the green light phase during the waiting period is D. min This indicates that within a green light cycle, vehicles from the target section to the target intersection can typically pass through the target intersection. Therefore, setting the target section here is to accurately define the range of traffic flow that can be fully released under the shortest green light duration, thus locking in the effective statistical area. Subsequent analysis of the number of vehicles passing through the target section and the duration of the monitoring segment yields the number of vehicles per unit passing through the target section, which can determine the traffic density of the lane, thereby providing a quantitative basis for dynamic signal light timing. The specific implementation is as follows: Divide the lane to be analyzed into a target section, which is a horizontal section taken from the lane at a distance equal to the target intersection. The section is perpendicular to the direction of vehicle travel. Deploy monitoring equipment at the target section and retrieve historical monitoring segments from the same time period as the current time. Based on the number of vehicles Q passing through the target section in the monitoring segment and the duration D of the monitoring segment, obtain the number of vehicles per unit passing through the target section N0 = Q / D.

[0028] In this embodiment, the unit of segment duration D is minutes, and the unit of vehicle number Q is the number of vehicles. Therefore, the unit vehicle number N0 represents the average number of vehicles passing through the target section within a unit minute.

[0029] Based on the steering ratio set, traffic function, and number of vehicles per unit, the reliable value of the green light phase to be scheduled is calculated for different timing durations. Based on the reliable value, the green light phase timing duration is automatically recommended, specifically: Get the allowed time range [D] min D max From the given time intervals, extract a specific time interval t, D. min ≤t≤D max Substituting the timing duration t of the green light phase to be scheduled into the traffic function, we obtain the first number of vehicles M. t ; The first number of vehicles is M. t It represents the number of vehicles expected to pass through the target intersection when the timing is t, considering the overall traffic flow duration and release efficiency. Using the end time of the immediately preceding red light phase of the green light phase to be scheduled as the statistical time, based on the surveillance video, count the number of vehicles N1 between the target section and the target intersection. Based on the number of vehicles per unit N0, calculate the expected number of vehicles N=N0(tD) when the green light phase to be scheduled has a duration of t. min )+N1,D minThis is the preset minimum timing duration; since N1 is the number of vehicles between the target section and the target intersection counted during the statistical time, from the perspective of the lane being analyzed heading towards the target intersection, these vehicles are the number of vehicles that can pass through the target intersection, while N0(tD) min The number of vehicles (N) represents the number of new vehicles that continuously enter the statistical interval from behind the target section and are able to pass smoothly to the intersection during the extra release period exceeding the minimum release time. Therefore, the expected number of vehicles (N) represents the number of vehicles that are expected to pass through the target intersection from the perspective of the existing traffic flow and the subsequent traffic flow.

[0030] Obtain the steering behavior probability of the lane to be analyzed in the green light phase at the time of matching from the set of steering ratios. Multiply the expected number of vehicles N by the corresponding steering behavior probability to obtain the second number of vehicles N. t Calculate the number N of the second vehicle. t This is because the vehicles counted in the projected number of vehicles N are not all heading towards the same turning behavior. For example, the green light at the time of matching may control vehicles to turn left, but only 20% of the vehicles in the lane being analyzed may be turning left. Therefore, it is necessary to multiply the projected number of vehicles N by the probability of the corresponding turning behavior to obtain the second number of vehicles N. t The second number of vehicles N here t It represents the number of vehicles expected to pass through the target intersection when the timing duration is t, based on the existing traffic flow and subsequent inbound traffic flow at the intersection. And calculate the target value X=N t / M t ; It should be noted that, since the first vehicle quantity M t The second number of vehicles N t These are the projected numbers of vehicles that can pass through the target intersection from different perspectives: the first is the number of vehicles M. t From the perspective of overall traffic flow duration and release efficiency, and the number of vehicles N t Considering both the existing congested traffic and the subsequent inflow of traffic at the intersection, a target value X that is too large or too small will cause problems. If the target value X is too large, it indicates that N... t Greater than M t This indicates that the current green light timing is too short, and the existing release efficiency cannot meet the actual traffic flow demand. This easily leads to lane congestion and road closures, hindering traffic flow. The target value X is too large, indicating that N... t Less than M t This indicates that the current green light timing is set too long, resulting in insufficient actual traffic flow. This will lead to idle and wasted green light resources at the intersection, reduce the overall traffic efficiency of the intersection, and prolong the waiting time of other phase lanes. Therefore, to ensure a reasonable duration, X should be set to N as much as possible. t / M t It is close to 1; see the following formula for details: The reliable value for calculating the timing duration t is Calculate the reliability values ​​corresponding to multiple timing durations, where each timing duration is within [D]. min D max Within this range, the timing duration corresponding to the maximum reliability value is taken as the recommended timing duration for the green light phase to be timed.

[0031] It should be noted that the formula Z is a function of X > 0, where Z ranges from 0 to 1. Specifically, when X is between 0 and 1, Z increases with increasing X; when X > 1, Z decreases with increasing X; and Z has a maximum value of 1 when X = 1. To achieve a reasonable duration setting, this application aims to maximize X = N. t / M t It's close to a 1 match because X=N t / M t The value of reliability Z is closer to 1. The larger the reliability Z value, the more reasonable the timing duration is. Therefore, it can be used as the recommended timing duration for the green light phase waiting to be matched.

[0032] It should be noted that this solution is based on real-time dynamic adjustment of traffic light timings according to traffic flow detection results, rather than simply relying on batch replacement of the entire timing scheme at fixed time intervals. Instead, it can flexibly adjust the red and green light durations cycle by cycle according to real-time changes in traffic flow at the intersection. For example, if the target intersection is a T-shaped intersection and the green light phases of each direction rotate in a predetermined order, and the green light from west to east was 50 seconds in the previous signal cycle, it may become 60 seconds in the next time cycle due to changes in traffic. This allows for cycle-by-cycle updates of timing instructions to address sudden changes in traffic flow, enabling real-time adjustment of the release strategy based on traffic flow status and achieving real-time response and dynamic control of traffic light timings.

[0033] This embodiment also provides a large model data collaborative management system for traffic, including a steering ratio set establishment module, a traffic function establishment module, a unit vehicle quantity calculation module, and a timing duration recommendation module; Steering ratio set establishment module: used to divide a day into multiple time windows using a fixed duration time interval division method; Match the current time to the time window, obtain the target intersection to be timed, and retrieve the historical monitoring segments of the target intersection within the time window; divide the target intersection area and the lane areas connected to the target intersection in the monitoring segments, and label the vehicles in the monitoring segments; analyze the turning behavior of vehicles at the target intersection, and establish a set of turning ratios for each lane at the target intersection. The passage function establishment module is used to obtain the number of vehicles passing through the target intersection and the passage time in historical surveillance videos, and to establish a passage function that changes the number of passing vehicles over time. Unit vehicle quantity calculation module: used to determine the current green light phase to be scheduled, take the passable lane controlled by the green light phase to be scheduled as the lane to be analyzed, calculate the target distance corresponding to the lane to be analyzed; determine the target section based on the target distance, and calculate the unit vehicle quantity passing through the target section based on the vehicles traveling on the lane to be analyzed. Timing recommendation module: It is used to calculate the reliable value of the green light phase to be scheduled for different timing durations based on the steering ratio set, traffic function and number of vehicles per unit, and automatically recommend the green light phase timing duration based on the reliable value.

[0034] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0035] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for collaborative management of large-scale model data in transportation, characterized in that, Comprising the following steps: Dividing one day into multiple time windows by means of a fixed-duration time interval division method; Matching the time window to which the current time belongs, acquiring the target intersection to be timing-adjusted, and retrieving historical monitoring segments of the target intersection within the time window; dividing the target intersection area in the monitoring segment and each lane area connected to the target intersection, marking vehicles in the monitoring segment, analyzing the turning behavior of vehicles at the target intersection, and establishing a turning ratio set for each lane at the target intersection; Acquiring the number of vehicles passing through the target intersection and the passing duration from historical monitoring videos, and establishing a passing function of the number of passing vehicles varying with time; Determining the current green phase to be timed, taking the passable lanes controlled by the green phase to be timed as lanes to be analyzed, and calculating the target distance corresponding to the lanes to be analyzed; Determining a target cross-section according to the target distance, and calculating the number of unit vehicles passing through the target cross-section based on vehicles traveling on the lanes to be analyzed; Calculating the reliability value when the green phase to be timed has different timing durations according to the turning ratio set, the passing function and the number of unit vehicles, and automatically recommending the timing duration of the green phase based on the reliability value.

2. The method for collaborative management of large-scale model data in transportation according to claim 1, characterized in that, Establishing the turning ratio set of each lane at the target intersection comprises: Dividing the target intersection area and each connected lane area in the monitoring segment, marking a vehicle C in the monitoring segment, wherein the vehicle C comes from lane L and passes through the target intersection to drive to another lane, dividing the area where the vehicle C is located by using a target detection algorithm, detecting a first direction when the vehicle C just touches the target intersection and a second direction when the vehicle C exits the target intersection, and determining the turning behavior of the vehicle C according to the first direction and the second direction; Determining the probability of each turning behavior according to the turning behavior of each vehicle that comes from lane L and passes through the target intersection to drive to other lanes, and thereby establishing the turning ratio set of lane L at the target intersection.

3. The method for collaborative management of large-scale model data for transportation according to claim 2, characterized in that, Establishing the passing function of the number of passing vehicles varying with time comprises: Extracting historical monitoring segments of the target intersection, acquiring the green cycle during which vehicles on lane L are allowed to pass and the red cycle during which vehicles are not allowed to pass, acquiring the number M of waiting vehicles to be driven that are on lane L during the red cycle, taking the vehicles that leave lane L during the next adjacent green cycle after the red cycle as passing vehicles, and recording the number of passing vehicles as m, where 0<m≤M, and recording the duration from the moment when the green light starts to the moment when the last passing vehicle leaves lane L as the passing duration; Determining multiple numbers of passing vehicles and multiple passing durations according to the corresponding multiple red cycles and green cycles in the monitoring segments, performing linear function fitting, and establishing the passing function of the number of passing vehicles varying with the passing duration.

4. The method for collaborative management of large-scale model data for traffic according to claim 1, characterized in that, Calculating the number of unit vehicles passing through the target cross-section comprises: Retrieving the lanes to be analyzed and the passable turning behaviors corresponding to the green phase to be timed, and calculating the target distance corresponding to the lanes to be analyzed according to the historical monitoring videos of the lanes to be analyzed; Divide the lane to be analyzed into a target section that is the same distance from the target intersection. Deploy monitoring equipment at the target section and retrieve historical monitoring segments that are consistent with the current time window. Based on the number of vehicles Q passing through the target section in the monitoring segment and the duration D of the monitoring segment, obtain the number of vehicles per unit passing through the target section N0 = Q / D.

5. The method for collaborative management of large-scale model data in transportation according to claim 4, characterized in that, Calculate the target distance corresponding to the lane to be analyzed, including: Obtain the allowable timing range of the green light phase to be matched [D] min D max ], where D min D is the preset minimum timing duration. max The preset maximum timing duration; Calculate the average speed of vehicles traveling in the lane to be analyzed, multiply it by the shortest timing duration, and obtain the target distance corresponding to the lane to be analyzed.

6. The method for collaborative management of large-scale model data for traffic according to claim 1, characterized in that, Automatically recommends green light phase timing duration based on reliability values, including: Substituting the timing duration t of the green light phase to be scheduled into the traffic function, we obtain the first number of vehicles M. t ; Using the end time of the immediately preceding red light phase of the green light phase to be scheduled as the statistical time, based on the surveillance video, count the number of vehicles N1 between the target section and the target intersection. Based on the number of vehicles per unit N0, calculate the expected number of vehicles N=N0(tD) when the green light phase to be scheduled has a duration of t. min )+N1,D min To determine the minimum preset timing duration, obtain the steering behavior probability of the lane to be analyzed during the green light phase of the desired timing from the set of steering ratios. Multiply the expected number of vehicles N by the corresponding steering behavior probability to obtain the second number of vehicles N. t And calculate the target value X=N t / M t ; The reliable value for calculating the timing duration t is Calculate the reliability values ​​corresponding to multiple timing durations, and take the timing duration corresponding to the maximum reliability value as the recommended timing duration for the green light phase to be timed.

7. A large-scale model data collaborative management system for traffic, used to execute the large-scale model data collaborative management method for traffic as described in any one of claims 1-6, characterized in that, The system includes a steering ratio set establishment module, a traffic function establishment module, a unit vehicle quantity calculation module, and a timing duration recommendation module; Steering ratio set establishment module: used to divide a day into multiple time windows using a fixed duration time interval division method; Match the current time to the time window, obtain the target intersection to be timed, and retrieve the historical monitoring segments of the target intersection within the time window; divide the target intersection area and the lane areas connected to the target intersection in the monitoring segments, and label the vehicles in the monitoring segments; analyze the turning behavior of vehicles at the target intersection, and establish a set of turning ratios for each lane at the target intersection. The passage function establishment module is used to obtain the number of vehicles passing through the target intersection and the passage time in historical surveillance videos, and to establish a passage function that changes the number of passing vehicles over time. Unit vehicle quantity calculation module: used to determine the current green light phase to be scheduled, take the passable lane controlled by the green light phase to be scheduled as the lane to be analyzed, and calculate the target distance corresponding to the lane to be analyzed; The target section is determined based on the target distance, and the number of vehicles per unit passing through the target section is calculated based on the vehicles traveling in the lane to be analyzed. Timing recommendation module: It is used to calculate the reliable value of the green light phase to be scheduled for different timing durations based on the steering ratio set, traffic function and number of vehicles per unit, and automatically recommend the green light phase timing duration based on the reliable value.

8. A large-scale model data collaborative management system for transportation according to claim 7, characterized in that, The steering ratio set establishment module includes a steering behavior determination unit and a steering ratio set establishment unit; Steering behavior determination unit: used to divide the target intersection area and the connected lane areas in the monitoring segment, label the vehicles in the monitoring segment, use the target detection algorithm to divide the area where the vehicle is located, detect the first direction when the vehicle just touches the target intersection and the second direction when it leaves the target intersection, and determine the steering behavior of the vehicle based on the first direction and the second direction. Steering Proportion Set Establishment Unit: Used to determine the probability of each steering behavior based on the steering behavior of each vehicle coming from the lane and passing through the target intersection to other lanes, and to establish the steering proportion set of the lane at the target intersection.

9. A large-scale model data collaborative management system for transportation according to claim 7, characterized in that, The unit vehicle quantity calculation module includes a target distance determination unit and a unit vehicle quantity calculation unit; Target distance determination unit: used to retrieve the lane to be analyzed and the passable turning behavior corresponding to the green light phase to be matched, and calculate the target distance corresponding to the lane to be analyzed based on the historical monitoring video of the lane to be analyzed; Unit vehicle quantity calculation unit: used to divide the target section within the lane to be analyzed, with a length equal to the target distance from the target intersection, deploy monitoring equipment at the target section, retrieve monitoring segments from the same historical time period that are consistent with the current time window, and obtain the unit vehicle quantity passing through the target section based on the number of vehicles passing through the target section in the monitoring segment and the duration of the monitoring segment.

Images