Green wave time period dynamic division and adaptive optimization method

Abstract

The invention discloses a green wave time period dynamic division and adaptive optimization method. The method comprises the following steps: acquiring a coordination phase and a corresponding coordination direction of each intersection in a line; performing intersection state judgment based on the historical flow data, and performing time period division according to different intersection states; screening intersections on the line in each time period to obtain a plurality of intersections as starting points of the green wave line; calculating the intersection signal period of each intersection in the line and the public period of the line, and determining the green light time of each phase of the intersection; identifying an overflow risk road section in the line, and performing green light time adjustment on a terminal intersection of the overflow risk road section; and performing intersection signal adjustment on the intersections in the green wave line according to the phase difference between the adjacent intersections. According to the method, the traffic states of different time periods are considered, the green light signal scheme of each time period is determined according to the specific time period, and adjustment is further performed according to the overflow state of the intersection, so that the traffic control scheme is more accurate, and the traffic efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a method for dynamic division and adaptive optimization of green wave time periods. Background Technology

[0002] With the increasing number of vehicles on urban roads, traffic efficiency has declined. Given the limited expansion of the road network, improving traffic management efficiency has become a crucial means of alleviating urban road congestion. Urban green wave traffic systems work by linking traffic lights along a road route so that when vehicles arrive at intersections, the traffic light for their direction of travel is green, allowing vehicles to pass through without stopping or with minimal stops. Existing green wave solutions primarily employ offline methods, optimizing traffic light control schemes based on historical data. However, in practical applications, the constantly changing characteristics of road traffic, such as significant differences in traffic volume at different times and the need for signal adjustments to alleviate traffic congestion when traffic overflows, mean that offline optimization schemes cannot adapt to dynamically changing traffic characteristics, thus failing to guarantee the traffic efficiency of green wave roads.

[0003] The Chinese patent document "Control Method, Device, Storage Medium, and Electronic Device for Traffic Signal Devices" (publication number CN118366323A, publication date July 19, 2024) includes: determining the common traffic cycle of each target intersection on a target road based on historical traffic data; determining the phase duration of each phase at each target intersection on the target road based on the common traffic cycle; determining the coordinated phase time difference between the target intersection and adjacent intersections; adjusting the forward coordinated phase time difference based on the reverse coordinated phase time difference to obtain the target forward coordinated phase time difference when a bidirectional green wave is allowed based on the reverse coordinated phase time difference; determining the target phase difference based on the reverse coordinated phase time difference and the target forward coordinated phase time difference; and controlling the traffic signal devices at the target intersection and adjacent intersections to operate according to the common traffic cycle, phase duration, and target phase difference. However, this technology also uses an offline traffic signal control scheme, and its control scheme based on historical data cannot adapt to dynamically changing traffic characteristics, thus failing to guarantee the traffic efficiency of green wave roads. Summary of the Invention

[0004] This invention aims to overcome the problem that most existing technologies use offline solutions based on historical data analysis to obtain traffic signal control results, which cannot adapt to the dynamically changing traffic operation characteristics on actual roads and cannot guarantee the traffic efficiency of green wave roads. It provides a method for dynamic division and adaptive optimization of green wave time periods.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for dynamic division and adaptive optimization of green wave time periods includes: Obtain the coordinated phase and corresponding coordinated direction at each intersection in the route; The intersection status is determined based on historical traffic data, and time periods are divided according to different intersection statuses; Within each time period, several intersections on the route are selected as the starting points for the green wave route. Calculate the intersection signal cycle of each intersection in the line and the common cycle of the line in each time period, and determine the green light time of each phase of the intersection based on the common cycle; Identify overflow risk sections in the route and adjust the green light time at the end intersection of the overflow risk section; For intersections in the green wave system, the intersection signals are adjusted based on the phase difference between adjacent intersections.

[0006] This invention first determines the coordinated phase and direction of each intersection as a reference for allocating more green light time during green wave route planning. Then, it assesses the intersection status based on historical traffic flow data, categorizing intersection status into different levels according to congestion severity. Based on these different status levels, the 24 hours of the day are divided into different traffic signal control periods. This allows for the application of different signal control schemes at different times based on actual traffic flow or congestion conditions, better adapting to dynamically changing traffic characteristics. For the selection of green wave routes, in addition to using the coordinated phase and direction of each intersection as a reference, it is also necessary to analyze specific needs or historical data to determine the starting intersections of the green wave routes, thus forming corresponding green wave routes for signal control. In the process of generating signal control schemes for each intersection, in addition to calculating the signal scheme based on the signal cycle and green light ratio of each intersection obtained from historical data, the schemes are also continuously updated and identified for overflow risk sections in the route. The signal schemes at the endpoints of the overflow risk sections are adjusted accordingly. This allows for fine-tuning of the signal control schemes based on actual congestion and traffic overflow at different times, better meeting the ever-changing traffic operation characteristics and improving traffic efficiency.

[0007] Preferably, the calculation of the intersection signal cycle of each intersection in the line and the common cycle of the line includes: identifying the overlapping phases and non-overlapping phases of the intersection, and calculating the minimum green light time of each phase of the intersection. Calculate the phase flow ratio of non-overlapping phases based on the flow ratio of each direction at the intersection; The intersection signal cycle is calculated based on the minimum green light time of each phase and the phase flow ratio of non-overlapping phases; the intersection signal cycle with the longest cycle in the line is selected as the common cycle.

[0008] As a preferred option, before calculating the intersection signal cycle of each intersection in the line, the phase sequence with the longest duration in the corresponding time period is selected as the intersection phase sequence for that time period. After completing the common cycle calculation of the route, for the last intersection in the coordinated direction, the phase sequence of the intersection during that time period is corrected according to the preset phase sequence judgment conditions.

[0009] Preferably, the determination of the green light time for each phase of the intersection based on the common cycle includes: Based on the intersection status in the coordinated direction according to the common cycle and the current time period, calculate the green light time for the coordinated phase and the non-coordinated phase respectively; If the sum of all green light times is greater than the common cycle, the excess time is subtracted from the non-coordinated phases where the green light time is greater than the minimum green light time of the phase; otherwise, the time difference is added to the coordinated phases on average. The green light time of the coordinated phase at adjacent intersections in the coordinated direction is corrected.

[0010] Preferably, the correction of the green light time for the coordinated phase of adjacent intersections in the coordinated direction includes: If the distance between two adjacent intersections in the coordination direction is greater than or equal to the first distance threshold, and the green light time of the upstream intersection minus the green light time of the downstream intersection is greater than the first time threshold, then the green light time of the downstream intersection is corrected. Filter out non-coordinated phases at downstream intersections whose green light time is longer than the minimum green light time of the phase, and add the excess time of the non-coordinated phases to the coordinated phases.

[0011] Preferably, the identified overflow risk sections in the route include: Obtain the hierarchy of all routes, and aggregate intersecting routes based on route aggregation conditions to obtain a route set; Iterate through the coordination directions of each segment in the route set, and calculate the overflow risk rate based on the difference between the flow leaving the segment and the flow entering the segment in the coordination direction. If the spillover risk rate is greater than the risk rate threshold, then there is an spillover risk in the coordination direction of the road segment.

[0012] Preferably, the line aggregation conditions include: Select the main route based on the subordinate relationship of the routes, and form a set R with all routes that intersect with the main route; If the difference between the cycle of a line in R and the cycle of the main line is greater than the cycle threshold, then the corresponding line is filtered out. Conversely, the common period of the corresponding line is set as the main line period, and the same judgment is made on the subsequent intersecting lines of the line. The lines that meet the conditions are added to R to obtain the aggregated line set. The remaining lines are obtained by filtering out the aggregated line sets. The remaining lines are then traversed and aggregated until no remaining lines are found.

[0013] Preferably, the adjustment of the green light time at the end intersection of the overflow risk road segment includes: Calculate the overflow parameter Δ for each flow direction j at the endpoint intersection. j and overflow correction time Δt j ; If the overflow parameters of one or two phases do not meet the preset conditions, calculate the overflow correction time Δt for the phase i that needs correction. i And the number of flow directions R that need to be corrected for this phase i ; For phase i that needs adjustment, add an overflow correction time Δt to the original green light time. i ; For phases that do not require adjustment, the increased green light time for phases that do require adjustment is shared evenly.

[0014] Preferably, the time period division based on different intersection conditions includes: Based on the historical traffic data of the intersection, set the tiered traffic range for the intersection, and the correspondence between the tiered traffic range and the intersection status; Set a unit interval time, and determine the intersection status at that unit interval time based on the stepped flow range where the average flow rate of each unit interval time is located. The time intervals of adjacent units with the same intersection status are merged to complete the time period division.

[0015] Preferably, obtaining the coordinated phase and corresponding coordinated direction of each intersection in the route includes: The entire day is divided into unit intervals, and each phase of the intersection is traversed to obtain the flow rate of each unit interval in the phase, thus obtaining the phase flow rate set. The maximum phase flow within the same unit interval is taken as the intersection flow within that unit interval, forming the maximum flow set of the intersection. The phase that appears most frequently in the maximum flow set is the coordination phase, and its direction is the coordination direction.

[0016] The present invention has the following beneficial effects: Based on the coordinated phase situation of each intersection, the time periods of each route are pre-divided, and the intersections that can serve as the starting point of the green wave in different time periods are determined. Then, the green light signal scheme for different time periods is determined, and the green light signal scheme is fine-tuned according to the overflow section to obtain the final green wave signal scheme. Compared with the prior art, it not only considers the traffic conditions of different time periods, but also determines the green light signal scheme for each time period according to the specific time period and further adjusts it according to the overflow state of the intersection, making the traffic control scheme more accurate and improving traffic efficiency. Attached Figure Description

[0017] Figure 1 This is a flowchart of the dynamic division and adaptive optimization method for green wave time periods in this invention. Detailed Implementation

[0018] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, a method for dynamic division and adaptive optimization of green wave time periods includes: Obtain the coordinated phase and corresponding coordinated direction at each intersection in the route; The intersection status is determined based on historical traffic data, and time periods are divided according to different intersection statuses; Within each time period, several intersections on the route are selected as the starting points for the green wave route. Calculate the intersection signal cycle of each intersection in the line and the common cycle of the line in each time period, and determine the green light time of each phase of the intersection based on the common cycle; Identify overflow risk sections in the route and adjust the green light time at the end intersection of the overflow risk section; For intersections in the green wave system, the intersection signals are adjusted based on the phase difference between adjacent intersections.

[0020] It should be noted that this invention first determines the coordinated phase and direction of each intersection as a reference for allocating more green light time during green wave route planning. Then, it judges the intersection status based on historical traffic flow data, classifying intersection status into different levels according to congestion severity. Based on these different intersection status levels, the 24 hours of the day are divided into different traffic signal control periods. This allows for the use of different signal control schemes at different times based on actual traffic flow or congestion, better adapting to dynamically changing traffic characteristics. Regarding the selection of green wave routes, in addition to using the coordinated phase and direction of each intersection as a reference, it is also necessary to analyze specific needs or historical data to determine the starting intersections of the green wave routes, thereby forming corresponding green wave routes for signal control. In the process of generating signal control schemes for each intersection, in addition to calculating the signal scheme based on the signal cycle and green light ratio of each intersection obtained from historical data, the schemes are also continuously updated and identified for overflow risk sections in the route. The signal schemes at the endpoints of the overflow risk sections are adjusted accordingly. This allows for fine-tuning of the signal control schemes based on actual congestion and traffic overflow at different times, better meeting the ever-changing traffic operation characteristics and improving traffic efficiency.

[0021] It is worth noting that for intersections in the green wave route, the intersection signal is adjusted based on the phase difference between adjacent intersections. That is, the phase difference between adjacent intersections is obtained by dividing the road segment distance between adjacent intersections by the recommended vehicle speed. With the signal control scheme at the starting intersection of the green wave route unchanged, the signal control scheme of the downstream intersections is delayed by the time corresponding to the phase difference. The specific method of green wave control based on the phase difference between adjacent intersections can be directly implemented using existing technologies, so it will not be described in detail.

[0022] Specifically, in actual traffic control, there exists a road network model that reflects actual road conditions. The location of each intersection can be obtained from the road network model, and different routes within the road network can be determined based on the actual situation. These different routes intersect to form a road network. The road network model can also be used to understand the distances between intersections, as well as the entrance / exit information, phase information, and lane-level flow direction information of each intersection. For vehicle-related traffic operation data, such as vehicle flow data in lane-level flow direction information and flow data in phase information, relevant acquisition devices can be used to collect and process the data, which is then correlated with the road network model. This allows the generation of various parameter data required for the dynamic division and adaptive optimization of green wave time periods in this invention.

[0023] In this invention, the coordinated phase and corresponding coordinated direction of each intersection in the route can be obtained by combining historical information with manual fine-tuning, or by analyzing and processing historical traffic data to obtain a more accurate coordinated phase and coordinated direction. It is worth noting that the coordinated phase and coordinated direction of each intersection in the route can be obtained in various ways, and using existing methods for determining the coordinated phase and coordinated direction does not affect the implementation of the subsequent methods in this invention.

[0024] After determining the coordinated phase and direction of each intersection, the intersection status is judged based on historical traffic data, and time periods are divided according to different intersection statuses. Intersection statuses are categorized into four types: oversaturated, congested, uncongested, and empty. Further subdivisions of specific intersection imbalance states can be made under congested and oversaturated conditions to obtain different intersection status results. Multiple adjacent time periods with the same intersection status are merged to obtain the time period division results for the entire day. It is worth noting that, in addition to the method mentioned in this invention, existing intersection status methods and time period division methods can also be used for judging intersection statuses and dividing time periods without affecting the subsequent acquisition of intersection signal control schemes.

[0025] After the time periods are defined, the intersection conditions differ at different times, resulting in different traffic conditions. Therefore, more targeted signal control schemes are needed for traffic control. This means generating corresponding signal control schemes for intersections in each time period (intersection signal cycle, green light ratio, and green light duration for each phase, etc.). This allows the control scheme to adapt to dynamically changing traffic characteristics and ensures road traffic efficiency.

[0026] The number of road segments suitable for green wave control varies depending on the time period. Some intersections can only be controlled as single-point intersections, while others can be controlled by forming green wave lines based on coordinated phase and direction. Therefore, all intersections on the line during each time period need to be screened to identify suitable starting points for green wave lines. Green wave coordination is then performed on these intersections and their subsequent intersections. Besides using the method of this invention, intersections selected as starting points for green wave lines can also be manually set or determined using existing methods without affecting the generation and adaptive optimization of the signal control scheme of this invention.

[0027] As a specific implementation, calculating the intersection signal cycle at each intersection in the line and the line's common cycle includes: Identify overlapping and non-overlapping phases at intersections and calculate the minimum green light time for each phase at the intersection. Calculate the phase flow ratio of non-overlapping phases based on the flow ratio of each direction at the intersection; The intersection signal cycle is calculated based on the minimum green light time of each phase and the phase flow ratio of non-overlapping phases; the intersection signal cycle with the longest cycle in the line is selected as the common cycle.

[0028] It should be noted that in this invention, the overlapping phase refers to a transitional state between two non-overlapping phases. For example, inserting an east-facing left-turn / straight-go phase between an east-west straight-ahead phase and an east-west left-turn phase is called an overlapping phase. When the traffic flow at the east and west entrances is mismatched, more vehicles can be allowed to pass through the east entrance. The minimum green light time for a phase is the same as the minimum green light time for that phase. In the specific calculation process, it is necessary to obtain road network model parameters, historical traffic flow data, previously obtained time period division results, and intersection characteristic parameters (intersection timing, phase sequence, and red / yellow light duration).

[0029] Specifically, the process of identifying overlapping phases at each intersection includes: Identify all flow directions at the intersection, traverse all phases of the intersection, identify the flow directions contained in other phases besides the current phase, if it is still all flow directions of the intersection, mark the current phase as an overlapping phase; after filtering out the overlapping phase, repeat the same filtering process from the remaining phases until all overlapping phases are filtered out, and the remaining phases are non-overlapping phases.

[0030] The calculation of the minimum green light time for each phase at the intersection is performed separately for overlapping and non-overlapping phases. If the phase is an overlapping phase in a non-coordinated direction, the minimum green light time is set to 0 seconds; if the phase is an overlapping phase in a coordinated direction, the minimum green light time is set to 12 seconds. For non-overlapping phases, the initial minimum green light time g is... min =12 seconds. If there are pedestrians crossing the street in this phase, calculate the minimum green light time for pedestrians to cross the street: In the formula, w is the lane width, typically taken as 3m, and s is the pedestrian walking speed, typically taken as 1m / s. n is the number of lanes corresponding to pedestrian crossings, determined by the pedestrian flow direction at the entrance corresponding to that phase. This invention only considers left-turn and straight-ahead phases, because right-turn phases are generally not controlled by traffic lights.

[0031] For the number of lanes n corresponding to pedestrian crossings, there are three cases for calculation.

[0032] The first type is the left-turn phase: When turning left from north or south, n = max(number of eastbound approach lanes + number of approach auxiliary lanes, number of westbound approach lanes + number of approach auxiliary lanes) + number of lanes corresponding to the green belt (if it exists, it is counted as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane); when turning left from east or west, n = max(number of southbound approach lanes + number of approach auxiliary lanes, number of northbound approach lanes + number of approach auxiliary lanes) + number of lanes corresponding to the green belt (if it exists, it is counted as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane).

[0033] The second scenario involves a straight-ahead phase, but the pedestrian does not have a second crossing: When traveling straight north and south, n = max(number of entrance lanes to the east + number of exit lanes + number of auxiliary road lanes, number of entrance lanes to the west + number of exit lanes + number of auxiliary road lanes) + number of lanes corresponding to the green belt (if it exists, it is counted as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane). When traveling straight east and west, n = max(number of southbound entrance lanes + number of exit lanes + number of auxiliary road lanes, number of northbound entrance lanes + number of exit lanes + number of auxiliary road lanes) + number of lanes corresponding to the green belt (if it exists, it is counted as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane).

[0034] The third type is the straight-ahead phase where pedestrians have two crossing options: For north-south straight traffic, n = max(number of lanes at the east entrance + number of lanes at the east auxiliary road, number of lanes at the east exit, number of lanes at the west entrance + number of lanes at the west auxiliary road, number of lanes at the west exit) + number of lanes corresponding to the green belt (if they exist, count as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane). For east-west straight traffic, n = max(number of lanes at the south entrance + number of lanes at the south auxiliary road, number of lanes at the south exit, number of lanes at the north entrance + number of lanes at the north auxiliary road, number of lanes at the north exit) + number of lanes corresponding to the green belt (if they exist, count as 1 lane) + number of lanes corresponding to the non-motorized vehicle lane (equivalent to 1 lane).

[0035] The intersection signal cycle is calculated based on the minimum green light time of each phase and the phase flow ratio of non-overlapping phases. First, the flow ratio x of each flow direction j at the intersection is calculated. j Perform the calculation: Where S j To achieve saturation flow, Q j S represents the average flow rate within the time period, where T is the duration of that time period (in hours); the time period consists of M hours, and S... m Let t be the saturation flow rate at hour m. m Let m be the duration of the time period in the m-th hour (for example, if the time period is 8:30-10:00, then M = 2, t1 = 30, t2 = 60, T = 1.5).

[0036] Based on the flow ratio of each flow direction j, the phase flow ratio of each phase at the intersection is then calculated: Step 1: Initialize the phase flux ratio y of each phase i. i =0, defines the flow direction as Copy_x j =x j Copy_mov i ={m i,1 ,…,m i,j If phase i includes flow direction j, then m i,j =1, otherwise m i,j =0.

[0037] Step 2: Based on Copy_mov i Identify all non-overlapping phases at the current intersection, and traverse all non-overlapping phases i: mark the flow direction combination J that is unique to phase i. Unique The flow direction that repeats with other phases is combined into J. Shared If J Unique If only pedestrian flow is considered, then If J Unique Including vehicle flow direction, then Deleting phase i is equivalent to clearing {m} to zero. i,1 ,…,m i,j} = 0, then for J Shared The flow direction j is updated sequentially with the flow direction ratio x. j =max(0,x j -y i Repeat step 2 for the remaining phase combinations to obtain the phase flow ratio for each non-overlapping phase.

[0038] When calculating the signal cycle at an intersection, the calculation can be categorized based on the completeness of the intersection data.

[0039] If the intersection data is complete, first calculate the equivalent phase number: I 普通 I represents the total number of non-overlapping phases. 搭接 To calculate the total number of overlapping phases, the intersection signal cycle is: Where w and Y are configurable parameters, they can be 1 and 0.8 respectively. The floor function rounds down to the nearest integer. i,min To calculate the minimum green light time for pedestrians crossing the street, A represents the yellow light time, and r represents the red light time. The yellow light time and red light time are determined by the initial timing of the intersection characteristic parameters.

[0040] In the absence of intersection data, the minimum cycle is calculated based on the minimum green light time and the number of phases for each phase: Among them, the yellow light T in each phase yellow and all red T red The timing scheme for this period is determined by obtaining the characteristic parameters of the intersection.

[0041] Calculate the maximum period based on the number of phases: The intersection size level is inferred based on the number of lanes N, and the average hourly lane flow rate is preset. The alternative signal cycle C is determined based on the average lane flow rate per hour. rec : The intersection signal cycle is C = max(T) fixed ,T min C rec ) Where T fixed T is the period of the timing scheme for the characteristic parameters of the intersection during the current time period. min The minimum common period is 130 by default; the round function means that when the last significant digit is 5 or greater, it rounds up, and when the last significant digit is less than 5, it rounds down.

[0042] The signal period at the intersection with the longest period in the line is taken as the common period C. R The period value is a multiple of 5 (rounded up).

[0043] Furthermore, before calculating the intersection signal cycle at each intersection in the line, the phase sequence with the longest duration in the corresponding time period is selected as the intersection phase sequence for that time period. After completing the common cycle calculation of the route, for the last intersection in the coordinated direction, the phase sequence of the intersection during that time period is corrected according to the preset phase sequence judgment conditions.

[0044] It should be noted that the last intersection for the coordinated direction is determined based on the results of the coordinated phase and coordinated direction. Once the coordinated phase and coordinated direction of an intersection have been determined, other coordinated directions will be disconnected at that intersection, meaning that this intersection is the last intersection for other coordinated directions.

[0045] Specifically, before calculating the intersection signal cycle for each intersection, the phase sequence with the longest duration is selected from the timing selection of intersection characteristic parameters as the intersection phase sequence for that time period. In addition to this phase sequence, other phase sequences are also included, collectively forming the phase sequence set P. n.

[0046] After completing the common cycle calculation of the route, for the last intersection in the coordinated direction, the phase sequence of the intersection during this period is corrected according to the preset phase sequence judgment conditions. For any two intersections, there is bidirectional traffic in both the up and down directions (except for one-way streets). Therefore, for the last intersection in the coordinated direction, it is necessary to consider both the up and down directions.

[0047] For the last intersection n in the uphill direction, obtain the set P of different phase sequences in its characteristic parameters. n ,calculate λ n =C R *rate in, These represent the travel time for the uphill and downhill segments at intersections n-1 and n, respectively. `rate` is a configuration parameter with a default value of 0.2.

[0048] Case 1: |x n |≥λ n Then there is no need to choose a better phase; Case 2: |x n |<λ n And x n >0, then in P n The phase sequence that prioritizes the downlink coordination phase is selected first. Case 3: |x n |<λ n And x n <0, then in P n The phase sequence that prioritizes the uplink coordination phase is selected first.

[0049] For the last intersection n in the downlink direction, obtain the set P of different phase sequences in its characteristic parameters. n ,calculate λ n =C R *rate in, These represent the travel time for the uphill and downhill segments at intersections n-1 and n, respectively. `rate` is a configuration parameter with a default value of 0.2.

[0050] Case 1: |x n |≥λ n Then there is no need to choose a better phase; Case 2: |x n |<λn And x n >0, then in P n The priority is to select the phase sequence that starts first for the uplink coordination phase and has the smallest difference from the previously determined phase sequence of the intersection phases for that time period. Case 3: |x n |<λ n And x n <0, then in P n The priority is to select the phase sequence that starts first for the downlink coordinated phase and has the smallest difference from the previously determined phase sequence of the intersection for that time period. The magnitude of the difference between the two phase sequences can be judged by the difference in green light time between them; the smallest difference in green light time indicates the smallest difference between the two phase sequences.

[0051] As a specific implementation, determining the green light time for each phase of an intersection based on a common cycle includes: Based on the intersection status in the coordinated direction according to the common cycle and the current time period, calculate the green light time for the coordinated phase and the non-coordinated phase respectively; If the sum of all green light times is greater than the common cycle, the excess time is subtracted from the non-coordinated phases where the green light time is greater than the minimum green light time of the phase; otherwise, the time difference is added to the coordinated phases on average. The green light time of the coordinated phase at adjacent intersections in the coordinated direction is corrected.

[0052] It should be noted that during signal timing calculations, if the duration of the common cycle is the minimum cycle, then green light time is not allocated based on the traffic flow ratio; the green light time for each phase is simply the minimum green light time for that phase. Furthermore, signal timing calculations also require consideration of whether the phase traffic flow ratio is 0 and the intersection status in the coordinating direction.

[0053] Phase flow rate y in each phase i When neither of the values ​​is 0, the initial phase green light time for phase i is: Among them, C R For the common cycle, A is the yellow light time; r is the red light time; g i,min y represents the minimum green light time for this phase. i Let y be the phase flow ratio, k be the number of phases, and ∑y be the sum of the flow ratios of each phase.

[0054] If the coordinated direction is congested or oversaturated during this period, the green light time for the non-coordinated phase is: g i =int(min(g) i ,max(g i,min Q i ×λi ))) The green light time for coordinated phases is: Where, λ i Configure parameters for the green light time of phase i, the default value is 2, Q i For phase flux, Q i =max(Q j ).

[0055] If the coordinated direction is empty or not congested during this period, the green light time for the non-coordinated phase is: The green light time for coordinated phases is: Where, λ i Configure parameters for the green light time of phase i, the default value is 2, Q i For phase flux, Q i =max(Q j If the current phase is a direct phase, then μ i =0.3, if this phase is a left-turning phase, then μ i =0.25, if this phase is a simultaneous release phase (all flows from one inlet are released at the same time), then μ i =0.15.

[0056] In the presence of arbitrary phases, the phase flow ratio y i When the green ratio is 0, the green ratio corresponding to the time period with the same weekday characteristics is taken. If there is no such period, the green ratio in the timing scheme of the intersection characteristic parameters is taken. The signal control scheme of the intersection is calculated based on the corresponding green ratio and signal cycle.

[0057] Green signal ratio λ of each phase in the coordination direction i for: If it is a non-overlapping phase, then If it is an overlapping phase, then γ i =0.1.

[0058] Green signal ratio λ in each phase of the non-coordinated direction i for: Where T fixed This refers to the period of the timing scheme for the intersection's characteristic parameters during the current time period.

[0059] After calculating the green light time for all phases at the intersection, verify that the sum of the green light times for all phases equals the common cycle. If the sum of the green light times is greater than the common cycle, determine if any phase in the non-coordinated phases has a green light time greater than the minimum green light time for that phase, and subtract the extra seconds from the average of these phases. If the sum of the green light times is less than the common cycle, add it to the average of the coordinated phases.

[0060] Furthermore, the green light time for the coordinated phase at adjacent intersections in the coordinated direction is corrected, including: If the distance between two adjacent intersections in the coordination direction is greater than or equal to the first distance threshold, and the green light time of the upstream intersection minus the green light time of the downstream intersection is greater than the first time threshold, then the green light time of the downstream intersection is corrected. Filter out non-coordinated phases at downstream intersections whose green light time is longer than the minimum green light time of the phase, and add the excess time of the non-coordinated phases to the coordinated phases.

[0061] It should be noted that the green light time correction for the coordinated phase of adjacent intersections in the coordinated direction is to coordinate the signal control of each intersection on the green wave line, so as to avoid the green light time and signal difference of adjacent intersections being too large, which would prevent green wave control from being achieved.

[0062] Specifically, after calculating the green light ratio at all intersections along the route, the green light times of the coordinated phase in the same coordinated direction are compared between adjacent intersections (i, i+1) in the coordinated direction. Taking the above row as an example, the comparison... and

[0063] Step 1: Determine whether the distance between intersection i and intersection i+1 is less than the first distance threshold of 700m. If yes, proceed to step 2; otherwise, return to continue comparing other adjacent intersections, intersection i+1 and i+2.

[0064] Step 2: Adjust the green light time at the downstream intersection if the following conditions are met: It is then assumed that the green light time at the downstream intersection i+1 needs to be adjusted. For the n non-coordinated phases at intersection i+1, calculate the maximum time that can be subtracted. That is, the green light time before phase x correction. Reduce its phase minimum green light time but For the n non-coordinated phases at intersection i+1, the green light time is subtracted from each of the following: Seconds (rounded up), minus the green light time of the last phase.

[0065] If the total time subtracted is less than min(10,ε), then If the phase is still greater than min(10,)-total, then subtract min(10,ε)-total from its green light time.

[0066] As a specific example, identifying overflow risk sections in the route includes: Obtain the hierarchy of all routes, and aggregate intersecting routes based on route aggregation conditions to obtain a route set; Iterate through the coordination directions of each segment in the route set, and calculate the overflow risk rate based on the difference between the flow leaving the segment and the flow entering the segment in the coordination direction. If the spillover risk rate is greater than the risk rate threshold, then there is an spillover risk in the coordination direction of the road segment.

[0067] Specifically, first obtain the line time period [t1, t2, ..., t n Get the 5-minute path traffic [Q1, Q2, ..., Q] of this line. 288 and path capacity C, if Q i If the temperature exceeds 85% C, the time period is marked as 1; otherwise, it is marked as 0. For each hourly segment, it is determined whether more than half of the 5 minutes are marked as 1. If so, that half-hour period is considered a congested period, and congestion control mode is activated. The above process is the determination process for whether congestion control is needed. Besides the method mentioned in this invention, existing congestion control triggering methods or manually setting corresponding thresholds can be used for determination and triggering.

[0068] Obtain the subordination relationships of all routes, and aggregate intersecting routes based on route aggregation conditions to obtain a route set; traverse the coordination direction of each segment in the route set, and calculate the overflow risk rate based on the difference between the flow leaving the segment and the flow entering the segment in the coordination direction.

[0069] For all aggregated routes, iterate through the road segments within each route (e.g., the road segment between adjacent intersections A and B) to calculate the coordinated direction. q in =q in,l +q in,r +q in,s q out =q out,l +q out,r +q out,s Where, q in,l q in,s q in,r These are the traffic flows entering the road segment from the left, straight, and right (the sum of these flows represents the total traffic flow entering that segment). If a three-way intersection lacks a flow direction, it can be removed.out,l q out,r q out,s The traffic flow corresponding to the exit segment (the sum of which represents the total traffic flow leaving that segment); t is the duration of the congestion period, l veh The length of the equivalent vehicle is 7m by default. L is the road segment length, r L This represents the spillover risk rate.

[0070] If r L If the value is greater than 0.7, then there is an overflow risk in that direction of the road segment. The direction of the corresponding road segment is marked as an overflow risk, and the end point of the overflow segment is marked as the overflow risk point.

[0071] Optionally, line aggregation conditions include: Select the main route based on the subordinate relationship of the routes, and form a set R with all routes that intersect with the main route; If the difference between the cycle of a line in R and the cycle of the main line is greater than the cycle threshold, then the corresponding line is filtered out. Conversely, the common period of the corresponding line is set as the main line period, and the same judgment is made on the subsequent intersecting lines of the line. The lines that meet the conditions are added to R to obtain the aggregated line set. The remaining lines are obtained by filtering out the aggregated line sets. The remaining lines are then traversed and aggregated until no remaining lines are found.

[0072] Specifically, obtain all route dependencies, prioritize the routes according to the number of times they intersect, and obtain a priority-sorted set; starting with the highest priority route, take it as the main route, and all routes that intersect with the group route together form a set R = {Route ...} i In this embodiment, the subordinate relationship of the routes is selected based on the road network model. It can be considered that it can be modified according to actual needs. For the priority sorting of the number of intersections, for example: if route A and route B intersect at intersection X, and route A and route C intersect at intersection Y, then the number of intersections of route A is 2. All routes are sorted in order of priority from large to small according to the number of intersections.

[0073] If the cycle of a line in set R differs from that of the main line by more than 10 seconds, the line is removed from set R and subsequently split and calculated based on the dividing point. If a line has multiple cycles spanning different time periods, the largest cycle value is taken as the line's common cycle. The common cycles of each line may differ; therefore, the common cycle is used as a criterion for determining whether aggregation is possible. If aggregation is not possible, the line is not added to set R, and the two lines need to be split at their intersection point. Subsequent calculations will then proceed after splitting the lines according to the dividing point.

[0074] If the difference between the cycle of a line in set R and the cycle of the main line is less than or equal to 10 seconds, the common cycle of the line is set as the cycle of the main line, the green light ratio is allocated according to the original scheme, the line is deleted from the priority sorting set, and the cycle of the subsequent intersecting lines of the line is judged to meet the condition that the difference is less than 10 seconds. If it meets the condition, the line is added to set R and the cycle of the line is updated to the cycle of the main line. Otherwise, the line is split and calculated according to the dividing point.

[0075] Remove all aggregated lines from the priority sorted set, and iterate through the remaining lines to perform the above line aggregation operation until all lines have been aggregated.

[0076] Furthermore, adjustments to the green light timing at the end intersections of road sections with overflow risk include: Calculate the overflow parameter Δ for each flow direction j at the endpoint intersection. j and overflow correction time Δt j ; If the overflow parameters of one or two phases do not meet the preset conditions, calculate the overflow correction time Δt for the phase i that needs correction. i And the number of flow directions R that need to be corrected for this phase i ; For phase i that needs adjustment, add an overflow correction time Δt to the original green light time. i ; For phases that do not require adjustment, the increased green light time for phases that do require adjustment is shared evenly.

[0077] Specifically, obtain the 5-minute average flow rate of each direction j at the endpoint intersection i of the overflow risk road segment. The length L of the upstream section flowing towards j j The red light waiting time t for direction j in the traffic signal control scheme before the intersection is revised. j (Signal cycle duration minus the green light time of all its phases).

[0078] Where ω is the segment length coefficient, which defaults to 0.7 (configurable), and Δ is the overflow parameter for flow direction j. j (Used to determine whether the road segment can handle the traffic flow) and overflow correction time Δt j .

[0079] When all flow directions satisfy Δ j If the value is ≥0, then the scheme does not need to be modified.

[0080] If one or two phases require correction (i.e., one or two phases have overflow parameters in the flow direction that are less than 0), then calculate the phase Δt that needs correction. i=max(Δt) j ), R i =len(phase) j (j represents the flow direction that needs an extended green light time), R i This indicates the number of flow directions that need to increase the green light time in phase i.

[0081] Phase according to x i Sort by size from largest to smallest, iterate through the phases that require an increase in green light time, and prioritize adjusting R. i The green light duration for a large phase. Phase green light duration t i =Δt i +t i For flow directions j with multiple phases, Δt is updated sequentially. j =max(0,Δt) j -Δt i ).

[0082] For n phases that do not require green light time adjustment, the increase in green light time is distributed evenly ∑ i Δt i That is, the green light time for each of the n phases is reduced respectively. If phase n is a non-overlapping phase, then its green light time is If phase n is an overlapping phase, then its green light time is

[0083] If the total subtracted time is less than ∑ i Δt i Then, priority should be given to finding the overlapping phase. Still greater than ∑ i Δt i -total phase, then subtract min(10,ε)-total from its green light time.

[0084] If there are more than two phases that need correction (i.e., more than two phases have overflow parameters of less than 0), no further optimization is needed, so no action is taken.

[0085] As an optional implementation, for each route in each aggregated route set R, the upstream intersection of the endpoint of the overflow risk section is marked as a stopping intersection, and the phase difference algorithm is called to solve the problem. Specifically, for the overflow risk section, if there is a one-way overflow, the inbound and outbound phases of the section are released simultaneously; if there is overflow in both directions, the phase difference is fixed at 0.5*T, where T is the common period (i.e., stopping upstream of the overflow section, simultaneous release of phases in the overflow section, and green wave downstream of the overflow section). After solving the phase difference for each route, taking the first route as the reference, the difference Δtp between the absolute phase differences obtained from the two routes at the intersection is calculated, and Δtp is uniformly added to other intersections of subsequent routes.

[0086] As a specific example, time period division based on different intersection conditions includes: Based on the historical traffic data of the intersection, set the tiered traffic range for the intersection, and the correspondence between the tiered traffic range and the intersection status; Set a unit interval time, and determine the intersection status at that unit interval time based on the stepped flow range where the average flow rate of each unit interval time is located. The time intervals of adjacent units with the same intersection status are merged to complete the time period division.

[0087] Specifically, in this embodiment, the time period is t, the flow direction is i, the phase is j, the inlet is m, the outlet is n, and the intersection-level 5-minute flow rate Q is used. t Flow-level 5-minute flow rate q t,i The input is the intersection model; configurable parameters include: three percentiles of the flow direction threshold (default 10%, 60%, 90%); three percentiles of the intersection threshold (default 10%, 60%, 90%); import flow imbalance index threshold UM_Th (default 1.5); same-direction import / export imbalance index threshold UE_Th (default 2); opposite-direction import imbalance index threshold UO_Th (default 2); and manually configured flow direction priority.

[0088] Taking the generation of a plan for any given day as an example, for any time period t, the equivalent average lane flow S in direction i t,i for: Calculate the average lane flow Q at the intersection t for: Where, n i Let λ be the number of lanes flowing towards i (using static information, ignoring dynamic information such as tidal flow and variable switching). i This is the flow direction correction factor (1 for straight lines, 1.2 for left turns or controlled right turns).

[0089] Calculate the flow direction threshold by taking the equivalent average lane flow S every five minutes over several days. t,i (Minimum 1 day, maximum 14 days) as a set Sort the elements of the set in ascending order. Use the 10th percentile value of the set (configurable) as the first flow threshold for flow direction i. Take the 60th percentile value of the set (configurable) as the second flow threshold for flow direction i: Take the 90th percentile value of the set (configurable) as the third flow threshold for flow direction i:

[0090] Calculate the intersection threshold and the intersection-level traffic flow Q over several days. t (Minimum 1 day, maximum 14 days) as a set Sort the elements of the set in ascending order, and take the 10th percentile value of the set (configurable) as the threshold for the first intersection: Use the 60th percentile value of the set (configurable) as the threshold for the second intersection: Use the 90th percentile value of the set (configurable) as the threshold for the third intersection: β1 is the threshold for the first intersection, with a default value of 4veh / 5min, and β2 is the threshold for the second intersection, with a default value of 14veh / 5min.

[0091] The algorithm inputs the recommended date for the current time by filtering similar dates from the previous month and calculating the average. Date categories are defined as follows: 1 - weekends or holidays, 2 - weekdays but the previous day is a weekend or holiday, 3 - weekdays but the next day is a weekend or holiday, and 4 - other weekdays. The algorithm extracts historical data from a set D of similar dates within the month preceding the recommended date, using the average value for that period as the algorithm input.

[0092] Where |D| represents the number of days of the same type of date in the set.

[0093] For each 5-minute segment, identify the intersection status and output a scene label. If there is no historical data, the default output time period is [0:00-7:00], [7:00-9:00], [9:00-12:00], [12:00-14:00], [14:00-16:45], [16:45-19:00], [19:00-23:59]. If historical data exists, first determine the intersection status.

[0094] The average lane flow rate at the intersection over 5 minutes should meet the following requirements: Or at least more than half of the flows should meet the following conditions. The overall rating is "oversaturated".

[0095] The average lane flow rate at the intersection over 5 minutes should meet the following requirements: Or all flows should satisfy The overall rating is "congested".

[0096] The average lane flow rate at the intersection over 5 minutes should meet the following requirements: Or (5-minute average lane flow at the intersection) But all flows are The overall rating is "non-congested".

[0097] The 5-minute average lane flow rate at the intersection meets the requirements. And at least one flow direction satisfies The overall rating is "empty".

[0098] In addition to assessing the intersection's condition as oversaturated, congested, uncongested, or vacant, it is also necessary to determine the intersection's imbalance status. The calculation of imbalance indicators includes: Calculation of Import Flow Imbalance Indicators (for Variable Lane Switching) For each import m, calculate the flow imbalance index. Where i∈m represents the flow direction i under inlet m.

[0099] Calculation of the imbalance index for import and export in the same direction (for tidal flow lane switching) Calculate the same-direction import / export imbalance index for each import / export group (m,n). Where i∈m represents the flow direction i under inlet m, and i∈n represents the flow direction i under outlet n.

[0100] Calculation of import imbalance index (phase sequence generation) Calculate the import imbalance index for each import pair (m1, m2). Where i∈m1 represents the flow direction i under inlet m1, and i∈m2 represents the flow direction i under outlet m2. If the order of the opposing inlet groups (m1,m2) is reversed, they need to be recalculated to ensure that the flow direction with the larger flow rate is listed first in the identified opposing unbalanced inlet groups.

[0101] Determining the imbalance at an intersection includes: Imbalance in import flow: Traverse each import m in the flow direction, if UM t,m >UM_Th, and the direction of the maximum flow rate under the inlet Record the import. The final output is a list of imports and flows with imbalanced flow {(m,i)}.

[0102] Imbalance between inbound and outbound traffic: Iterate through each inbound / outbound group (m,n) in the flow direction, if UE t,(m,n) >UE_Th, and the direction of the maximum flow rate under the inlet Record the import. The final output is a list of imports and exports with an imbalance {(m,n)}.

[0103] Imbalance of opposing inlets: Traverse each opposing inlet group (m1, m2) under the flow direction, if Furthermore, the m1 inlet meets the requirements for both downward and leftward flow. Record the import. The final output is a list of imports with a corresponding import imbalance {(m1, m2)}.

[0104] The intersection status, the import-flow list of imbalanced flow, the import-export list of imbalanced import and export, and the import list of imbalanced opposite imports are integrated into a scene tag L. t Obtain the label sequence for each five minutes within a day from the aforementioned steps, i.e., {L0,…,L...} t ,…,L 288}

[0105] During the time period segmentation process, the intersection states are merged. The first time period is initialized to [00:00, 00:05]. Each time point is iterated. If the intersection filling time is the same as the previous time point, the start time remains unchanged, and the duration is continuously incremented. If they are different, the end of the previous scene is recorded, along with the start time, end time, and duration of the previous scene. The current scene at the intersection is designated as the new scene time period, with the start time being the end time of the previous scene and the duration being 5 minutes.

[0106] The process iterates through oversaturated, congested, uncongested, and idle scenarios in sequence, removing excessively short time periods from highest to lowest (first iterating through saturated scenarios, then sequentially iterating through congested, uncongested, and idle scenarios). Taking saturation as an example, initialize t... last Given 100000000000, iterate through each time period. If it's a saturated scene, if tt last <10 minutes (the end time of a time period minus the start time, i.e., the duration of the time period is less than 10 minutes), adjust t. last The time period up to time t is labeled as saturated; and the end time t of the most recent saturated scene is recorded. last =t.

[0107] Iterate through all time periods. If a time period lasts less than 15 minutes, adjust the end time of the preceding period and the start time of the following period to the midpoint of that time period (rounded up to the nearest 5 minutes), and then delete that time period. Output the overall intersection status time periods as [t0,…,t…]. k ,…,t K ] After completing the merging of intersection states, further perform time period merging based on the imbalance state of the intersection. For the time period merging of any same type of events, define the following process, and merge according to the flow priority configured manually. Calculate the import priority based on the sum of the flow priorities under each import. Traverse the imbalance event time periods of each import respectively, delete the imbalance time periods with an imbalance duration less than 15 minutes, and correct them to no imbalance occurred.

[0108] Initialize the time period [t0, …, t k , …, t K as the imbalance event time period of the import with the highest priority, read the imbalance event time period [T0, …, T k′ , …, T K of the next priority import, and arrange the two time periods in chronological order as [t0, T0, T1, t1, t2, T2, t3, …, t n , 23:59]. If the time difference between two adjacent times t k1 and T k2 is less than 15 minutes, then merge the time periods divided by the algorithm into the regular time period t k1 = T k2 .

[0109] For example: T1 < t1 < t2 < T2 < t3, and the time interval between the two time periods t2 and T2 is less than 15 minutes, then divide the time periods into [00:00, T1], [T1, t1], [t1, T2], [T2, t3]. Then the time periods [00:00, T1], [T1, t1] and [00:00, t1] have the same scenario, the time period [t1, T2] has the same scenario as the time period [t1, t2], the time period [T2, t3] has the same scenario as the time period [t2, t3], and so on.

[0110] Traverse all time periods [t k , t k:1 in turn. If the intersection state is empty or non-congested, record the signal time period as [t k , t k:1 . If the intersection state is congested or oversaturated, traverse the opposite-import imbalance events and merge them according to the priority, and record the signal time period as [t k , …, t p , …, t k:1 and the opposite-import imbalance events corresponding to the scenario. The variable lanes and tidal lanes are also merged according to the priority for the time periods.

[0111] As another optional embodiment, the intersection status can be divided into five states: oversaturation, saturation, congestion, non-congestion, and empty. In this embodiment, saturation and congestion correspond to a further subdivision of the congestion state among the four intersection states in the previous embodiment. Therefore, the oversaturation, saturation, and congestion states in this embodiment can be processed according to the oversaturation and congestion states in the aforementioned embodiment. Similarly, the non-congestion and empty states in this embodiment can be processed according to the non-congestion and empty states in the aforementioned embodiment.

[0112] Specifically, the calculation involves the status of the route and its intersections within the specified time period: The flow rate S every five minutes over several days will be converted to the average lane flow rate. t,i (Minimum 1 day, maximum 14 days) as a set Sort the elements of the set in ascending order. Use the 20th percentile value of the set (configurable) as the first flow threshold for flow direction i. Take the 50th percentile value of the set (configurable) as the second flow threshold for flow direction i: Take the 70th percentile value of the set (configurable) as the third flow threshold for flow direction i: Take the 90th percentile value of the set (configurable) as the fourth flow threshold for flow direction i:

[0113] The coordination direction of several days for import-level flow Q t (Minimum 1 day, maximum 14 days) as a set Sort the elements of the set in ascending order, and take the 10th percentile value of the set (configurable) as the first inflow threshold: Take the 50th percentile value of the set (configurable) as the second inlet flow threshold: Take the 70th percentile value of the set (configurable) as the third inlet flow threshold: Take the 90th percentile value of the set (configurable) as the fourth inlet flow threshold: The default value for β1 is 6 veh / 5 min, the default value for β2 is 10 veh / 5 min, the default value for β3 is 16 veh / 5 min, and the default value for β4 is 25 veh / 5 min.

[0114] Iterate through each time period and determine if it exceeds 30 minutes. If so, divide the time period into 30-minute intervals. For example, time period T... n [08:10~09:20] can be divided into [08:10, 08:40, 09:10, 09:20]. If the last time interval is less than 15 minutes, it is merged with the previous time interval, i.e., T′ n [08:10, 08:40, 09:20]. Regarding T′ n For each hourly segment within the timeframe, calculate the status of the inlet in the coordination direction. For example, for the north-south coordination direction, the inlet is the south inlet and the north inlet.

[0115] The 5-minute average flow rate of the inlet should meet the following requirements: Or all flows should satisfy The overall rating is "oversaturated". If the length of the upstream road section of the intersection is greater than 1km and the number of lanes of the upstream exit lane is ≥2, then the speed limit for the upstream road section in the coordinated direction of the intersection is set at 35km / h; otherwise, the speed limit for the upstream road section in the coordinated direction of the intersection is set at 30km / h.

[0116] The 5-minute average flow rate of the inlet should meet the following requirements: Or all flows should satisfy The overall rating is "saturated". If the length of the upstream road section of the entrance is greater than 1km and the number of lanes of the upstream exit lane is ≥2, then the speed limit of the upstream road section in the coordinated direction of the intersection is set at 45km / h; otherwise, the speed limit of the upstream road section in the coordinated direction of the intersection is set at 40km / h.

[0117] The 5-minute average flow rate of the inlet should meet the following requirements: Or all flows should satisfy The overall rating is "Congested". If the length of the upstream road section of the entrance is greater than 1km and the number of lanes of the upstream exit lane is ≥2, then the speed limit for the upstream road section in the coordinated direction of the intersection is set at 55km / h; otherwise, the speed limit for the upstream road section in the coordinated direction of the intersection is set at 50km / h.

[0118] The 5-minute average flow rate of the inlet should meet the following requirements: Or all flows should satisfy The overall rating is "non-congested". If the length of the upstream road section of the entrance is greater than 1km and the number of lanes of the upstream exit lane is ≥2, then the speed limit for the upstream road section in the coordinated direction of the intersection is set at 65km / h; otherwise, the speed limit for the upstream road section in the coordinated direction of the intersection is set at 60km / h.

[0119] The imported 5-minute average flow rate meets And at least one flow direction satisfies The overall rating is "empty". If the length of the upstream road section of the entrance is greater than 1km and the number of lanes of the upstream exit lane is ≥2, then the speed limit for the upstream section in the coordinated direction of the intersection is set at 75km / h; otherwise, the speed limit for the upstream section in the coordinated direction of the intersection is set at 70km / h.

[0120] The speed of the road segment determined under different conditions is used as the recommended speed for vehicles in the green wave line for the phase difference calculation of the green wave line.

[0121] As a specific implementation, obtaining the coordinated phase and corresponding coordinated direction of each intersection in the line includes: dividing the whole day into unit interval time, traversing each phase of the intersection, and obtaining the flow rate of each unit interval time in the phase to obtain the phase flow rate set; The maximum phase flow within the same unit interval is taken as the intersection flow within that unit interval, forming the maximum flow set of the intersection. The phase that appears most frequently in the maximum flow set is the coordinating phase, and its direction is the coordinating direction. Taking a green wave with a north-south direction as an example, the phase containing the north-south straight flow direction is the coordinating phase.

[0122] Specifically, if the intersection has several coordinated phases, and some phases have missing traffic flow data, then the default allocation method will be used. The vehicle capacity of the entrance where coordinated phase i is located will be calculated. Where j is the inlet j where the coordinated phase i is located; L ij Where is the length of the road segment at entrance j; v is the equivalent length of a vehicle, defaulting to 7m; N ij Let J be the number of lanes on the road segment leading to entrance j. Compare the C values ​​for each coordinated phase. i Choose C i Coordinate directions in the smaller direction, and disconnect other directions at the intersection.

[0123] For intersections with complete intersection data, take the 5-minute traffic flow for each direction and calculate the set of 5-minute traffic flows for the entire day for coordinated phase i. in, k represents the k-th flow direction in phase i; that is, the five-minute flow rate with the largest flow rate in phase i.

[0124] Determine the maximum phase flow in each 5-minute period to obtain the set. in, The phase with the highest flow rate during time period t is j. The judgment set R... 路口 The phase that appears most frequently in the intersection is the direction in which that phase is located, and that direction is the coordinated direction of the intersection.

[0125] For other coordinated directions, determine whether the direction is disconnected at the intersection or merged into a sub-area for coordination.

[0126] Taking an east-west coordinated phase as an example, we determine whether north-south left turns at the intersection require coordination. For southbound traffic, we assess the flow rates for both straight and left turns. We then determine the maximum flow rate for each phase over a 5-minute period to obtain a set. in, The phase with the highest flow rate during time period t is j. The judgment set R... 路口 The phase with the most frequent turns, if it is a straight-ahead phase, then the road breaks off at that intersection; if it is a left-turn phase, then the intersection to the south and the east-west roads are merged into a sub-zone, and the coordinating direction at that intersection is left turn. The same applies to northbound roads.

[0127] As a specific implementation example, a method is also provided for selecting several intersections on the route as the starting point of the green wave route within each time period. Specifically, Step 1: Obtain the distance between two adjacent intersections on the main road. Filter by the distance between upstream and downstream intersections. i l j Intersections with all distances greater than L are sorted by spacing, and the spacing (l) is determined by the distance between them. i +l j Larger intersections are prioritized. A recommended intersection sequence [Z1, Z2, ...] is obtained, with the recommendation reason tag for intersections in the sequence being: "Intersection spacing too large." Here, L is a configuration parameter, with a default value of 1000m.

[0128] Step 2: Obtain the average 5-minute flow rate at each intersection within the time period. For each intersection, calculate the average flow rate q1 of the coordinated direction and the average flow rate q2 of the non-coordinated direction with the largest flow rate. The algorithm filters intersections where r is greater than threshold R1 (configurable, default 2) and q2 is greater than threshold Q1 (the 70th percentile of the 5-minute flow rate for that direction). It then sorts the filtered intersections according to the value of r, placing those with larger r values ​​first, resulting in a ratio list and a recommended intersection sequence [B1, B2, ...]. The recommendation reason for the intersections in the sequence is labeled as: the average flow rate of the non-coordinated direction lane is greater than that of the coordinated direction lane.

[0129] Step 3: Let E be the recommended list of line cutting points. 。 Initialize the flag for each intersection i =0, used to indicate the recommended score for this intersection.

[0130] Step 3.1: Determine if the intersection recommendation sequence [Z1, Z2, ...] is empty. If empty, no processing is required; if not empty, add it to the final recommendation list E. Since there is a distinction between uphill and downhill traffic within a road segment, such as a green wave in the north-south direction, south to north can be defined as uphill and north to south as downhill. Therefore, there are separate final recommendation lists E for uphill and downhill directions. up or Edown Taking the upward direction as an example, for E already added to the recommendation list... up Intersection Notes

[0131] Step 3.2: Determine if the recommended intersection sequence [B1, B2, ...] is empty. If it is empty, no processing is required; if it is not empty, add the intersections in the sequence to E in sequence. up E up =[Z 11 ..., B1, ...], and modify the flag. i +=r i .

[0132] Step 3.3: Determine if there are 9 or more consecutive intersections in any direction without recommended intersections. If so, select a compromise number of intersections *i* as parking points to ensure the number of intersections on the sub-route remains within a preset range (e.g., 4-7). Modify the flags of these intersections. i += 100, and add it to E up .

[0133] Step 4: For the final recommendation list E up Intersection screening is conducted.

[0134] Step 4.1: For intersections with an upstream road section less than 300m, from E... up Remove from the list (except for manually selected points, which are either manually designated or directly used as algorithm input).

[0135] Step 4.2: Traverse all intersections in the path. If there is more than one recommended parking intersection among any three consecutive intersections, compare their flags. i Only keep the flag i The largest intersection, if flag i If they are the same, only the upstream intersection will be retained (except for manually selected points and the first intersection in that direction, such as the first intersection in the south-to-north direction in the green wave line).

[0136] Step 5: Output the line cutting point and the reason for recommendation. Cut the line, and the cutting point is the starting point of the next green wave line segment.

[0137] The above embodiments are further elaborations and descriptions of the present invention to facilitate understanding, and are not intended to limit the present invention in any way. Any modifications, equivalent substitutions, and improvements 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 dynamic division and adaptive optimization of green wave time periods, characterized in that, include: Obtain the coordinated phase and corresponding coordinated direction at each intersection in the route; The intersection status is determined based on historical traffic data, and time periods are divided according to different intersection statuses; Within each time period, several intersections on the route are selected as the starting points for the green wave route. The intersection signal cycle and the common cycle of the line are calculated for each intersection in each time period, and the green light time for each phase of the intersection is determined based on the common cycle. Identify overflow risk sections in the route and adjust the green light time at the end intersection of the overflow risk section; For intersections in the green wave system, the intersection signals are adjusted based on the phase difference between adjacent intersections.

2. The method for dynamic division and adaptive optimization of green wave time periods according to claim 1, characterized in that, The calculation includes the intersection signal cycle at each intersection in the route and the common cycle of the route, including: Identify overlapping and non-overlapping phases at intersections and calculate the minimum green light time for each phase at the intersection. Calculate the phase flow ratio of non-overlapping phases based on the flow ratio of each direction at the intersection; The intersection signal cycle is calculated based on the minimum green light time of each phase and the phase flow ratio of non-overlapping phases. The signal cycle at the intersection with the longest cycle in the line is selected as the common cycle.

3. The method for dynamic division and adaptive optimization of green wave time periods according to claim 1 or 2, characterized in that, Before calculating the intersection signal cycle at each intersection in the line, the phase sequence with the longest duration in the corresponding time period is selected as the intersection phase sequence for that time period. After completing the common cycle calculation of the route, for the last intersection in the coordinated direction, the phase sequence of the intersection during that time period is corrected according to the preset phase sequence judgment conditions.

4. The method for dynamic division and adaptive optimization of green wave time periods according to claim 1 or 2, characterized in that, The determination of the green light time for each phase of the intersection based on the common cycle includes: Based on the intersection status in the coordinated direction according to the common cycle and the current time period, calculate the green light time for the coordinated phase and the non-coordinated phase respectively; If the sum of all green light times is greater than the common cycle, the excess time is subtracted from the non-coordinated phases where the green light time is greater than the minimum green light time of the phase; otherwise, the time difference is added to the coordinated phases on average. The green light time of the coordinated phase at adjacent intersections in the coordinated direction is corrected.

5. The method for dynamic division and adaptive optimization of green wave time periods according to claim 4, characterized in that, The correction of the green light time for the coordinated phase of adjacent intersections in the coordinated direction includes: If the distance between two adjacent intersections in the coordination direction is greater than or equal to the first distance threshold, and the green light time of the upstream intersection minus the green light time of the downstream intersection is greater than the first time threshold, then the green light time of the downstream intersection is corrected. Filter out non-coordinated phases at downstream intersections whose green light time is longer than the minimum green light time of the phase, and add the excess time of the non-coordinated phases to the coordinated phases.

6. A method for dynamic division and adaptive optimization of green wave time periods according to claim 1, 2, or 5, characterized in that, The identified overflow risk sections in the route include: Obtain the hierarchy of all routes, and aggregate intersecting routes based on route aggregation conditions to obtain a route set; Iterate through the coordination directions of each segment in the route set, and calculate the overflow risk rate based on the difference between the flow leaving the segment and the flow entering the segment in the coordination direction. If the spillover risk rate is greater than the risk rate threshold, then there is an spillover risk in the coordination direction of the road segment.

7. The method for dynamic division and adaptive optimization of green wave time periods according to claim 6, characterized in that, The line aggregation conditions include: Select the main route based on the subordinate relationship of the routes, and form a set R with all routes that intersect with the main route; If the difference between the cycle of a line in R and the cycle of the main line is greater than the cycle threshold, then the corresponding line is filtered out. Conversely, the common period of the corresponding line is set as the main line period, and the same judgment is made on the subsequent intersecting lines of the line. The lines that meet the conditions are added to R to obtain the aggregated line set. The remaining lines are obtained by filtering out the aggregated line sets. The remaining lines are then traversed and aggregated until no remaining lines are found.

8. A method for dynamic division and adaptive optimization of green wave time periods according to claim 1, 2, 5, or 7, characterized in that, The adjustment of green light time at the end intersection of the overflow risk road section includes: Calculate the overflow parameter Δ for each flow direction j at the endpoint intersection. j and overflow correction time Δt j ; If the overflow parameters of one or two phases do not meet the preset conditions, calculate the overflow correction time Δt for the phase i that needs correction. i And the number of flow directions R that need to be corrected for this phase i ; For phase i that needs adjustment, add an overflow correction time Δt to the original green light time. i ; For phases that do not require adjustment, the increased green light time for phases that do require adjustment is shared evenly.

9. A method for dynamic division and adaptive optimization of green wave time periods according to claim 1, 2, 5, or 7, characterized in that, The time period division based on different intersection conditions includes: Based on the historical traffic data of the intersection, set the tiered traffic range for the intersection, and the correspondence between the tiered traffic range and the intersection status; Set a unit interval time, and determine the intersection status at that unit interval time based on the stepped flow range where the average flow rate of each unit interval time is located. The time intervals of adjacent units with the same intersection status are merged to complete the time period division.

10. A method for dynamic division and adaptive optimization of green wave time periods according to claim 1, 2, 5, or 7, characterized in that, The acquisition of the coordinated phase and corresponding coordinated direction of each intersection in the route includes: The entire day is divided into unit intervals, and each phase of the intersection is traversed to obtain the flow rate of each unit interval in the phase, thus obtaining the phase flow rate set. The maximum phase flow within the same unit interval is taken as the intersection flow within that unit interval, forming the maximum flow set of the intersection. The phase that appears most frequently in the maximum flow set is the coordination phase, and its direction is the coordination direction.

Images