A green-to-total ratio optimization method based on adaptive control

Abstract

The application discloses a green signal ratio optimization method based on adaptive control. The method periodically collects traffic flow data and dynamically adjusts the green signal ratio of the traffic signal by using an adaptive control algorithm. According to the traffic congestion conditions of different time periods and different road sections, the green signal ratio can be accurately and efficiently optimized, the road traffic capacity can be effectively improved, the average waiting time of vehicles at the intersection can be reduced, and the traffic congestion degree can be reduced. Compared with the traditional fixed green signal ratio setting mode, the method significantly improves the flexibility and intelligence of traffic signal control, and provides an innovative solution for efficient operation of urban traffic.

Description

Technical Field

[0001] This invention belongs to the field of intelligent traffic control technology, and in particular, it is a green ratio optimization method based on adaptive control. Background Technology

[0002] With the acceleration of urbanization, traffic congestion has become a common problem faced by major cities worldwide. Traffic congestion not only increases travel time costs but also leads to energy waste, environmental pollution, and frequent traffic accidents. Statistics show that the economic losses caused by urban traffic congestion account for 1%-3% of a city's GDP. As an important tool for urban traffic management, optimizing traffic light control strategies is crucial for alleviating traffic congestion.

[0003] The green light ratio refers to the ratio of green light time to the cycle length within a signal period. A well-designed green light ratio can balance traffic demand from different directions, improve intersection efficiency, and reduce vehicle delays and queue lengths. However, traditional green light ratio optimization methods are mostly based on fixed cycles and static traffic flow data, making it difficult to adapt to real-time changes in traffic conditions. This is especially true during peak hours or special events (such as exams or large-scale events), where the dynamic and complex nature of traffic flow makes the limitations of traditional methods even more apparent.

[0004] Traditional green light ratio optimization methods (such as the Webster model) are primarily applicable to unsaturated traffic conditions, with the optimization objective of minimizing delays and stops. However, these methods often perform poorly under saturated or congested traffic conditions, and may even worsen traffic conditions. For example, the Webster model cannot effectively handle traffic flow differences between phases under saturated traffic conditions, leading to a continuous increase in queue lengths in some directions, which in turn causes more severe congestion. Furthermore, traditional methods lack the ability to dynamically respond to real-time traffic data and are ill-suited to adapting to rapid changes in traffic flow.

[0005] Wu Peng et al. (Wu Peng, Ye Baolin, Wu Weimin, Chen Bin, Zhang Yijia. Traffic Signal Control Based on Improved Chaotic Particle Swarm Algorithm. Acta Metrologica Sinica, 2024, 45(12): 1876-1884.) proposed a signal optimization method based on the improved Chaotic Particle Swarm Algorithm (ICPSO). The ICPSO algorithm introduces a neighborhood radius parameter for elite particles with high fitness in the population, and implements neighborhood chaotic search, which can improve the ability of particles to escape local optima while retaining their advantageous characteristics. Simulation experiments show that this method is better than the signal control model using a fixed signal period.

[0006] Luo Haocheng et al. (Luo Haocheng, Ren Bin, He Chunhong. Traffic signal dual-layer synchronous optimization control based on improved genetic algorithm [J]. Journal of Dongguan University of Technology, 2025, 32(1):35-40.) constructed a traffic signal dual-layer synchronous optimization control model and improved the genetic algorithm. In the dual-layer model, the upper layer optimizes the phase sequence according to the corresponding constraints, and the lower layer optimizes the traffic signal timing according to the results of the upper layer. The genetic algorithm adopts an elite strategy and an adaptive mutation operator, which are used to retain the optimal traffic signal control scheme and improve the optimization efficiency, respectively. The method has adaptive capability and high convergence efficiency, and can effectively reduce the average vehicle waiting time and improve the vehicle diversion efficiency through the SUMO traffic simulation experimental platform.

[0007] Wen Feng et al. (Wen Feng, Zhang Kexin. Research on Traffic Signal Timing Optimization Based on Deep Reinforcement Learning [J]. Journal of Shenyang University of Technology, 2019, 38(1): 48-52.) proposed a traffic signal timing optimization method based on deep reinforcement learning. On the basis of traditional Q-learning, a deep reinforcement learning strategy is adopted to optimize the signal timing technology at intersections, so as to reduce the number of vehicles in the traffic system and the average travel time of vehicles through intersections, thereby improving the efficiency of the traffic system.

[0008] Shen Guoqing (Shen Guoqing. Research on Traffic Light Timing Optimization Technology Based on 3D Convolutional Deep Reinforcement Learning [D]. Shenyang University of Technology, 2021. DOI:10.27323 / d.cnki.gsgyc.2021.000415.) proposed a traffic light timing method based on convolutional deep reinforcement learning. This method uses a convolutional neural network model to simulate the characteristics of the traffic network in the spatiotemporal dimension. It inputs standardized spatiotemporal dynamic road network information data and uses deep reinforcement learning algorithms to analyze the road network control actions at intersections. Through continuous feedback learning, it constructs a dynamic control system for intersection traffic lights that can deeply reinforce accumulated experience, enabling intelligent control of intersection traffic lights.

[0009] Abu et al. (Jamil ARM, Ganguly KK, Nower N. Adaptive traffic signal control system using composite reward architecture based deep reinforcement learning[J]. IET Intelligent Transport Systems, 2020, 14(14): 2030-2041.) proposed an adaptive control method based on deep reinforcement learning. Considering that the agent's feedback depends on the reward function, the authors proposed a novel reward architecture called Composite Reward Architecture (CRA) for multi-objective adaptive traffic signal control to optimize multiple objectives. It computes multiple rewards in parallel for each action and uses majority voting to select the desired action.

[0010] Mustafa et al. (Coşkun M, Baggag A, Chawla S. Deep reinforcement learning for traffic light optimization[C] / / 2018 IEEE International Conference on DataMining Workshops (ICDMW). IEEE, 2018: 564-571.) proposed a signal optimization method based on deep reinforcement learning. This method designs a novel reward function that simultaneously considers traffic flow and delay. This reward function is applied to the adaptive traffic control problem and maximized using deep learning methods. Experiments use deep Q-learning (DQN) and deep policy gradient (DPG) as benchmark deep reinforcement learning methods, and combine them with SUMO to verify the usability of the method.

[0011] The above methods can be categorized into expert-experience-based methods, rule-based methods, heuristic algorithm-based methods, and deep reinforcement learning-based methods. Expert-experience-based methods require experts to input a large amount of configuration information into the signal system, meaning they cannot adapt to continuously changing traffic flows and have weak adaptability. Heuristic algorithm-based green ratio optimization methods often require defining a fitness function and continuously searching the solution space. The performance of heuristic algorithms is often highly dependent on specific problem characteristics and parameter settings, and they are prone to getting trapped in local optima. While some heuristic algorithms can theoretically find good solutions, they may require significant computational resources and time in practical applications. Deep reinforcement learning-based methods require obtaining reward values ​​and training separately for different intersections. In real-world applications, many parameters affect the reward value. This type of method struggles to find a universal action and state space, and the extensive training and prediction processes consume massive amounts of server computing resources, making it difficult to apply to current urban road networks composed of hundreds or thousands of intersections. Summary of the Invention

[0012] The purpose of this invention is to address the problems existing in the prior art by providing a green ratio optimization method based on adaptive control, which can be applied to signal controllers or deployed in intelligent traffic signal control systems for central-level control of signal controllers.

[0013] The technical solution to achieve the purpose of this invention is: a green light ratio optimization method based on adaptive control. Specifically, the method uses the queue length and release time of each stage in the previous cycle as input, and generates the release time of each stage in the next cycle before the end of the previous cycle. The cycle is defined as the sum of the time required for all traffic signal phases to complete one cycle. The stage is defined as each change of right-of-way at the intersection within a cycle. The phase is defined as a group of traffic flows that simultaneously acquire right-of-way within a signal cycle.

[0014] Furthermore, the method includes the following steps:

[0015] Step 1: Convert the queue length of each phase into the queue length of the stage, and further convert it into the stage adjustment time;

[0016] Step 2: Redistribute the stage adjustment time to each stage according to the time constraints to obtain the stage time for each stage;

[0017] Step 3: Send the generated stage time to the signal machine for execution.

[0018] Further, step 1, which involves converting the queue length of each phase into the queue length of a stage, and then further converting it into a stage adjustment time, specifically includes:

[0019] Step 1-1: Obtain the set of queue lengths for all stages in the (T-1)th cycle after the green light ends. ;gather for ,in This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i.

[0020] Steps 1-2 define the set of sequence numbers for stages with fixed release times as follows: , This represents the sequence number of the Mth stage where the release time is fixed, where M is the total number of sequence numbers in set F;

[0021] Define the set of sequence numbers of the stages whose maximum queue length in the previous cycle was 0 as follows: , This represents the sequence number of the Lth stage where the maximum queue length is 0, where L is the total number of sequence numbers in set U.

[0022] After removing F and U from the set of all stage sequence numbers, the set of stage sequence numbers for which release time needs to be optimized is obtained. , Indicates the first The sequence number of the stage where the release time needs to be optimized. Let X be the total number of indices in set X, and N be the total number of stages in one cycle.

[0023] Steps 1-3, based on sets F, U, and X respectively, divide all stages of the (T-1)th cycle into fixed sets of stages. Set of skippable stages and the set of stages that require optimized release times ;

[0024] Steps 1-4, Calculation The average total queue length in each stage :

[0025]

[0026] In the formula, This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i.

[0027] Steps 1-5, Calculation Queue length and time at each stage The difference:

[0028]

[0029] In the formula, The stage number is The queuing length and time of the stage The difference; The stage number is indicated as The maximum queue length after the green light ends in all phases included in the T-1th cycle;

[0030] Steps 1-6, define the sequence number as The stage with the largest difference is the stage where the difference is largest. Adjust step size accordingly Then calculate the step size ratio. for:

[0031]

[0032] Steps 1-7, according to calculate Adjustment time for all stages:

[0033]

[0034] In the formula, for The serial number is The adjustment period for the current stage.

[0035] Furthermore, step 2 specifically includes:

[0036] Step 2-1: Based on the stage adjustment time obtained in Step 1, obtain the Tth cycle stage. phase time :

[0037]

[0038] In the formula, It is the Tth cycle stage The phase time; It is the T-1th cycle stage. Release time;

[0039] Step 2-2, based on the stage time obtained in Step 2-1, according to Time constraints, generation A new phase and adjustment time :

[0040]

[0041]

[0042] Seek all The sum of the two is as follows ;

[0043] In the formula, and They represent the serial numbers respectively. The maximum and minimum stage times of the phase;

[0044] Steps 2-3, Calculation Adjustment time for all stages The sum of the two is as follows ;

[0045] Steps 2-4, taking into account The total time (c) of not releasing passengers at all stages is recorded.

[0046]

[0047] In the formula, This represents the stage time of stage i in the Tth cycle;

[0048] Steps 2-5, based on the new phase time Total adjustment time The time for each stage is allocated to different stages, resulting in the final stage time for each stage.

[0049] Furthermore, steps 2-5 specifically include:

[0050] like Then proceed to step 3, if Then proceed to steps 2-6, if Then proceed to steps 2-7;

[0051] Steps 2-6: Take 1 second from A and allocate this 1 second to... stage Above, among which The index k is determined by adding the first element of set F to each element of set X, and then sorting the elements in ascending order, denoted as k. The specific process of this step includes:

[0052] Take them in order of k Execute in the following order:

[0053] Let j=1;

[0054] Step 2-6-1, k is taken ;

[0055] Step 2-6-2: Allocate the extracted 1 second to the stage. Above, determine this stage If the maximum stage time is reached after adding 1 second, proceed to step 2-6-3; otherwise, proceed to step 2-6-4.

[0056] Step 2-6-3: Determine if A is 0. If yes, proceed to step 3; otherwise, execute: Take another second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-6-1. If k does not reach its maximum value, then return to step 2-6-1.

[0057] Step 2-6-4: Let j = j + 1, then return and re-execute step 2-6-1;

[0058] Steps 2-7: Take -1 second from A and allocate this -1 second to... stage Above; the specific process of this step includes:

[0059] Take them in order of k Execute in the following order:

[0060] Let j=1;

[0061] Step 2-7-1, k is taken ;

[0062] Step 2-7-2: Allocate the extracted -1 second to the stage. Above, determine this stage If the minimum stage time is reached after adding or subtracting 1 second, proceed to step 2-7-3; otherwise, proceed to step 2-7-4.

[0063] Step 2-7-3: Determine if A is 0. If yes, proceed to step 3; otherwise, execute: Take -1 second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-7-1. If k does not reach its maximum value, then return to step 2-7-1.

[0064] Step 2-7-4: Let j = j + 1, then return and re-execute step 2-7-1.

[0065] On the other hand, a green ratio optimization system based on adaptive control is provided, the system comprising:

[0066] The first module is used to convert the queue length of each phase into the queue length of the stage, and further into the stage adjustment time;

[0067] The second module is used to redistribute the stage adjustment time to each stage according to the time constraints, and obtain the stage time for each stage.

[0068] The third module is used to send the generated stage time to the signal machine for execution.

[0069] On the other hand, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the adaptive control-based green ratio optimization method when executing the computer program.

[0070] On the other hand, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the green ratio optimization method based on adaptive control.

[0071] Compared with the prior art, the significant advantages of this invention are:

[0072] (1) This invention generates a scheme for each intersection in real time on a periodic basis, mainly optimizing the release time at each stage, i.e., the green wave ratio. Furthermore, during the optimization of the green wave ratio, this invention can also ensure that the designed green wave effect does not deteriorate.

[0073] (2) This invention introduces a fixed phase, which firstly solves the requirement that the coordination effect cannot change, and secondly satisfies the requirement that the release time of a certain phase cannot change when there is no detector in that phase. To ensure coordination, the coordination phase can be placed in the first phase and the first phase can be added to the fixed phase. In this way, under a fixed period, the coordination effect remains unchanged even if the time of other phases changes.

[0074] (3) The present invention also takes into account the safety requirements in practical applications, such as the introduction of maximum stage time and minimum stage time. In the process of generating the final stage time, it is determined whether the release time of each stage meets the requirements. For stages that do not meet the requirements, the stage time is first adjusted to the range of maximum stage time and minimum stage time. Then, in order to ensure that the cycle remains unchanged, the adjusted time is redistributed to other stages according to the time constraints. On the basis of ensuring safety, the queue length and stage time are matched to the maximum extent.

[0075] (4) This invention is a rule-based signal optimization method, which has the following advantages compared with other existing methods: It does not rely on the complex model design of heuristic methods, and the calculation of intersection schemes can be completed in milliseconds. It does not require experts to perform a lot of pre-setting and statistical work for each intersection, saving manpower costs; Compared with deep reinforcement learning methods, its advantages are more obvious. First, this invention is a universally applicable method, which does not require a large number of servers to train the parameters and traffic flow information of each intersection, thus saving a lot of resources. Only one server is needed to generate schemes for hundreds of signal controllers; second, deep reinforcement learning methods are also difficult to meet the requirements of introducing safety parameters.

[0076] The present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0077] Figure 1 This is a flowchart of a green ratio optimization method based on adaptive control.

[0078] Figure 2 This is a logic diagram for allocating excess time in one embodiment.

[0079] Figure 3 This is a missing time allocation logic diagram in one embodiment. Detailed Implementation

[0080] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0081] It should be noted that if the embodiments of the present invention involve descriptions such as "first" and "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" and "second" may explicitly or implicitly include at least one of those features. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0082] In one embodiment, an adaptive control-based green light ratio optimization method is provided. Specifically, the method uses the queue length and clearance time of each stage in the previous cycle as input to generate the clearance time for each stage of the next cycle before the previous cycle ends. The cycle is defined as the sum of the time required for all traffic signal phases to complete one cycle. The stage is defined as each change of right-of-way at the intersection within a cycle. The phase is defined as a group of traffic flows that simultaneously acquire right-of-way within a signal cycle.

[0083] Furthermore, in one embodiment, combined with Figure 1 The method includes the following steps:

[0084] Step 1: Convert the queue length of each phase into the queue length of the stage, and further convert it into the stage adjustment time;

[0085] Step 2: Redistribute the stage adjustment time to each stage according to the time constraints to obtain the stage time for each stage;

[0086] Step 3: Send the generated stage time to the signal machine for execution.

[0087] Preferably, in some embodiments, step 1, which converts the queue length of each phase into the queue length of the stage, and further into the stage adjustment time, specifically includes:

[0088] Step 1-1: Obtain the set of queue lengths for all stages in the (T-1)th cycle after the green light ends. ;gather for ,in This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i.

[0089] Here, a set of sequence numbers for stages with fixed time is introduced, which can include the sequence numbers of the coordination stage and the stage without bound detectors.

[0090] Steps 1-2 define the set of sequence numbers for stages with fixed release times as follows: , This represents the sequence number of the Mth stage where the release time is fixed, where M is the total number of sequence numbers in set F;

[0091] Define the set of sequence numbers of the stages whose maximum queue length in the previous cycle was 0 as follows: , This represents the sequence number of the Lth stage where the maximum queue length is 0, where L is the total number of sequence numbers in set U.

[0092] After removing F and U from the set of all stage sequence numbers, the set of stage sequence numbers for which release time needs to be optimized is obtained. , Indicates the first The sequence number of the stage where the release time needs to be optimized. Let X be the total number of indices in set X, and N be the total number of stages in one cycle.

[0093] Steps 1-3, based on sets F, U, and X respectively, divide all stages of the (T-1)th cycle into fixed sets of stages. Set of skippable stages and the set of stages that require optimized release times ;

[0094] Steps 1-4, Calculation The average total queue length in each stage :

[0095]

[0096] In the formula, This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i.

[0097] Steps 1-5, Calculation Queue length and time at each stage The difference:

[0098]

[0099] In the formula, The stage number is The queuing length and time of the stage The difference; The stage number is indicated as The maximum queue length after the green light ends in all phases included in the T-1th cycle;

[0100] Steps 1-6, define the sequence number as The stage with the largest difference is the stage where the difference is largest. Adjust step size accordingly Then calculate the step size ratio. for:

[0101]

[0102] Steps 1-7, according to calculate Adjustment time for all stages:

[0103]

[0104] In the formula, for The serial number is The adjustment period for the current stage.

[0105] Here, the queue length at each stage is compared with the average, and an adjustment step size is introduced to convert the difference into demand adjustment time.

[0106] Here, this invention first collects the queue length after each phase ends and is released in real time in the previous cycle. For an intersection, an excellent adaptive algorithm needs to balance the traffic flow in all directions, allocating more time to phases with longer queue lengths and less time to phases with shorter queue lengths. Regarding the definition of queue length, this invention chooses the queue length after the phase ends and is released, which is more practically meaningful, rather than the maximum queue length within the cycle. This is because the maximum queue length gradually decreases as traffic is released, while the queue length after the phase ends and is released reflects whether the traffic flow and green light time are matched.

[0107] In order to prevent the release of vehicles in stages where there is no demand, this invention also needs to collect the maximum queue length of each phase in the previous cycle at the end of the previous cycle. If the maximum queue length of all phases in a stage is 0, then the stage can be skipped to prevent the green light from being released in vain.

[0108] This invention first converts the queue length of each phase into the queue length of each stage, and then compares the queue length of each stage with the average. An adjustment step size is introduced to convert the difference into a demand adjustment time. This completes the transformation from queue length to time. Therefore, stages with longer queue lengths will ultimately have increased time, while stages with shorter queue lengths will ultimately have decreased time, which aligns with our optimization objective. During the generation of stage times, since the demand adjustment time uses the queue length of the previous cycle, the final demand adjustment time is added to the release time of the previous cycle.

[0109] Preferably, in some embodiments, step 2 specifically includes:

[0110] Step 2-1: Based on the stage adjustment time obtained in Step 1, obtain the Tth cycle stage. phase time :

[0111]

[0112] In the formula, It is the Tth cycle stage The phase time; It is the T-1th cycle stage. Release time;

[0113] Step 2-2, based on the stage time obtained in Step 2-1, according to Time constraints, generation A new phase and adjustment time :

[0114]

[0115]

[0116] Seek all The sum of the two is as follows ;

[0117] In the formula, and They represent the serial numbers respectively. The maximum and minimum stage times of the phase;

[0118] Steps 2-3, Calculation Adjustment time for all stages The sum of the two is as follows ;

[0119] Steps 2-4, taking into account The total time (c) of not releasing passengers at all stages is recorded.

[0120]

[0121] In the formula, This represents the stage time of stage i in the Tth cycle;

[0122] Steps 2-5, based on the new phase time Total adjustment time The time for each stage is allocated to different stages, resulting in the final stage time for each stage.

[0123] Preferably, in some embodiments, steps 2-5 specifically include:

[0124] like Then proceed to step 3, if Then proceed to steps 2-6, if Then proceed to steps 2-7;

[0125] Steps 2-6, combined Figure 2 Take 1 second from A and allocate this 1 second to stage Above, among which The index k is determined by adding the first element of set F to each element of set X, and then sorting the elements in ascending order, denoted as k. The specific process of this step includes:

[0126] Take them in order of k Execute in the following order:

[0127] Let j=1;

[0128] Step 2-6-1, k is taken ;

[0129] Step 2-6-2: Allocate the extracted 1 second to the stage. Above, determine this stage If the maximum stage time is reached after adding 1 second, proceed to step 2-6-3; otherwise, proceed to step 2-6-4.

[0130] Step 2-6-3: Determine if A is 0. If yes, proceed to step 3; otherwise, execute: Take another second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-6-1. If k does not reach its maximum value, then return to step 2-6-1.

[0131] Step 2-6-4: Let j = j + 1, then return and re-execute step 2-6-1;

[0132] Steps 2-7, combined Figure 3 Take -1 second from A and allocate this -1 second to stage Above; the specific process of this step includes:

[0133] Take them in order of k Execute in the following order:

[0134] Let j=1;

[0135] Step 2-7-1, k is taken ;

[0136] Step 2-7-2: Allocate the extracted -1 second to the stage. Above, determine this stage If the minimum stage time is reached after adding or subtracting 1 second, proceed to step 2-7-3; otherwise, proceed to step 2-7-4.

[0137] Step 2-7-3: Determine if A is 0. If so, proceed to step 3 (send the generated stage time to the signal machine for execution, where the stage time in set F is equal to the stage time of the previous cycle, the stage in set U is not allowed, and the stage time in set X is allowed). If not, then execute: take -1 second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-7-1. If k does not reach its maximum value, then return to step 2-7-1.

[0138] Step 2-7-4: Let j = j + 1, then return and re-execute step 2-7-1.

[0139] Here, the maximum stage time and minimum stage time are introduced. During the generation of stage time, it is determined whether the release time of each stage meets the requirements. For stages that do not meet the requirements, these stage times need to be adjusted to the range of the maximum stage time and minimum stage time.

[0140] In one embodiment, a green ratio optimization system based on adaptive control is provided, the system comprising:

[0141] The first module is used to convert the queue length of each phase into the queue length of the stage, and further into the stage adjustment time;

[0142] The second module is used to redistribute the stage adjustment time to each stage according to the time constraints, and obtain the stage time for each stage.

[0143] The third module is used to send the generated stage time to the signal machine for execution.

[0144] Specific limitations regarding the green ratio optimization system based on adaptive control can be found in the limitations of the green ratio optimization method based on adaptive control mentioned above, and will not be repeated here. Each module in the aforementioned green ratio optimization system based on adaptive control can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0145] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements:

[0146] Step 1: Convert the queue length of each phase into the queue length of the stage, and further convert it into the stage adjustment time;

[0147] Step 2: Redistribute the stage adjustment time to each stage according to the time constraints to obtain the stage time for each stage;

[0148] Step 3: Send the generated stage time to the signal machine for execution.

[0149] For specific limitations on each step, please refer to the limitations on the green ratio optimization method based on adaptive control mentioned above, which will not be repeated here.

[0150] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program being implemented when executed by a processor:

[0151] Step 1: Convert the queue length of each phase into the queue length of the stage, and further convert it into the stage adjustment time;

[0152] Step 2: Redistribute the stage adjustment time to each stage according to the time constraints to obtain the stage time for each stage;

[0153] Step 3: Send the generated stage time to the signal machine for execution.

[0154] For specific limitations on each step, please refer to the limitations on the green ratio optimization method based on adaptive control mentioned above, which will not be repeated here.

[0155] As a specific example, the invention will be further verified and illustrated in one embodiment.

[0156] This invention proposes a green ratio optimization method based on adaptive control, which specifically includes the following process:

[0157] Step 1-1, the cycle has 4 stages. The maximum stage time is defined as 80 seconds, and the minimum stage time is defined as 15 seconds. This includes the east-west straight-ahead phase, where the maximum queue length is 30 meters. This includes the east-west left-turn phase, so the maximum queue length obtained by default is 0; The north-south straight phase has a maximum queue length of 10 meters. This is the North-South left-turn phase. The maximum queue length during the North-South left-turn phase is 20 meters. Therefore, the queue lengths for each phase of the previous cycle... for The set of all stages of the previous period (T-1) is: .

[0158] Steps 1-3, among which This is a coordination phase, therefore the phase duration cannot be changed. , ; The phase includes undeployed detectors, i.e. , And can be obtained , .

[0159] Steps 1-4, The average total queue length in each stage .

[0160] Steps 1-5, Calculation Queue length and time at each stage The difference: , .

[0161] Steps 1-6, The stage with the largest difference, the difference in this stage. Adjust step size accordingly Then calculate the step size ratio. .

[0162] Steps 1-7, according to calculate Adjustment time for all stages , .

[0163] Step 2-1: Based on the stage adjustment time obtained in Step 1, obtain the stage time of the Tth cycle. ,

[0164] Step 2-2: Based on the stage time obtained in Step 2-1, and according to the time constraints of the stage, the following can be calculated: , , , ;

[0165] Seeking ;

[0166] Steps 2-3, Calculation The sum of adjustment times for all stages ;

[0167] Steps 2-4, taking into account The total time spent with no passengers allowed to pass through any stage of the process is recorded. ;

[0168] Steps 2-5, based on the new phase time Total adjustment time The time for each stage is allocated to different stages, resulting in the final stage time for each stage.

[0169] because Distribute 10 seconds sequentially to and During the allocation process and The phase time is within the maximum and minimum time range, therefore and Each gets 5 seconds, therefore The final stage lasts 20 seconds. The final stage lasts 35 seconds. The duration of each stage is as follows: The cycle is 85 seconds, which is consistent with the previous cycle time and meets the requirement of keeping the coordination cycle unchanged.

[0170] In summary, the method of this invention dynamically adjusts the green ratio of traffic signals by periodically collecting traffic flow data and employing an adaptive control algorithm. It can accurately and efficiently optimize the green ratio based on traffic congestion conditions at different times and on different road sections, effectively improving road capacity, reducing the average waiting time of vehicles at intersections, and alleviating traffic congestion. Compared to traditional fixed green ratio settings, this method significantly improves the flexibility and intelligence of traffic signal control, providing an innovative solution for the efficient operation of urban traffic.

[0171] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention without departing from its spirit and scope should be included within the protection scope of the present invention.

Claims

1. A green ratio optimization method based on adaptive control, characterized in that, The method takes the queue length and clearance time of each stage in the previous cycle as input, and generates the clearance time of each stage in the next cycle before the end of the previous cycle. The cycle is defined as the sum of the time required for all traffic signal phases to be displayed for one cycle. The stage is defined as each change of right-of-way at the intersection within a cycle. The phase is defined as a group of traffic flows that simultaneously acquire right-of-way within a signal cycle. The method includes the following steps: Step 1: Convert the queue length of each phase into the queue length of the stage, and further convert it into the stage adjustment time; Step 2: Redistribute the stage adjustment time to each stage according to the time constraints to obtain the stage time for each stage; Step 3: Send the generated stage time to the signal machine for execution; Step 1, which involves converting the queue length of each phase into the queue length of the stage, and further converting it into the stage adjustment time, specifically includes: Step 1-1: Obtain the set of queue lengths for all stages in the (T-1)th cycle after the green light ends. ;gather for ,in This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i. Steps 1-2 define the set of sequence numbers for stages with fixed release times as follows: , This represents the sequence number of the Mth stage where the release time is fixed, where M is the total number of sequence numbers in set F; Define the set of sequence numbers of the stages whose maximum queue length in the previous cycle was 0 as follows: , This represents the sequence number of the Lth stage where the maximum queue length is 0, where L is the total number of sequence numbers in set U. After removing F and U from the set of all stage sequence numbers, the set of stage sequence numbers for which release time needs to be optimized is obtained. , Indicates the first The sequence number of the stage where the release time needs to be optimized. Let X be the total number of indices in set X, and N be the total number of stages in one cycle. Steps 1-3, based on sets F, U, and X respectively, divide all stages of the (T-1)th cycle into fixed sets of stages. Set of skippable stages and the set of stages that require optimized release times ; Steps 1-4, Calculation The average total queue length in each stage : In the formula, This represents the maximum queue length after the green light ends in the (T-1)th cycle for all phases included in phase number i. Steps 1-5, Calculation Queue length and time at each stage The difference: In the formula, The stage number is The queuing length and time of the stage The difference; The stage number is indicated as The maximum queue length after the green light ends in all phases included in the T-1th cycle; Steps 1-6, define the sequence number as The stage with the largest difference is the stage where the difference is largest. Adjust step size accordingly Then calculate the step size ratio. for: Steps 1-7, according to calculate Adjustment time for all stages: In the formula, for The serial number is The adjustment period for each stage; Step 2 specifically includes: Step 2-1: Based on the stage adjustment time obtained in Step 1, obtain the Tth cycle stage. phase time : In the formula, It is the Tth cycle stage The phase time; It is the T-1th cycle stage. Release time; Step 2-2, based on the stage time obtained in Step 2-1, according to Time constraints, generation A new phase and adjustment time : Seek all The sum of the two is as follows ; In the formula, and They represent the serial numbers respectively. The maximum and minimum stage times of the phase; Steps 2-3, Calculation Adjustment time for all stages The sum of the two is as follows ; Steps 2-4, taking into account The total time (c) of not releasing passengers at all stages is calculated. In the formula, This represents the stage time of stage i in the Tth cycle; Steps 2-5, based on the new phase time Total adjustment time The time is allocated to each stage, resulting in the final stage time for each stage.

2. The green ratio optimization method based on adaptive control according to claim 1, characterized in that, Steps 2-5 specifically include: like Then proceed to step 3, if Then proceed to steps 2-6, if Then proceed to steps 2-7; Steps 2-6: Take 1 second from A and allocate this 1 second to... stage Above, among which The index k is determined by adding the first element of set F to each element of set X, and then sorting the elements in ascending order, denoted as k. The specific process of this step includes: Take them in order of k Execute in the following order: Let j=1; Step 2-6-1, k is taken ; Step 2-6-2: Allocate the extracted 1 second to the stage. Above, determine this stage If the maximum stage time is reached after adding 1 second, proceed to step 2-6-3; otherwise, proceed to step 2-6-4. Step 2-6-3: Determine if A is 0. If yes, proceed to step 3; otherwise, execute: Take another second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-6-1. If k does not reach its maximum value, then return to step 2-6-1. Step 2-6-4: Let j = j + 1, then return and re-execute step 2-6-1; Steps 2-7: Take -1 second from A and allocate this -1 second to... stage Above; the specific process of this step includes: Take them in order of k Execute in the following order: Let j=1; Step 2-7-1, k is taken ; Step 2-7-2: Allocate the extracted -1 second to the stage. Above, determine this stage If the minimum stage time is reached after adding or subtracting 1 second, proceed to step 2-7-3; otherwise, proceed to step 2-7-4. Step 2-7-3: Determine if A is 0. If yes, proceed to step 3; otherwise, execute: Take -1 second from A, set j = j + 1, and determine if k has reached its maximum value. If k reaches its maximum value, then set j=1 and return to step 2-7-1. If k does not reach its maximum value, then return to step 2-7-1. Step 2-7-4: Let j = j + 1, then return and re-execute step 2-7-1.

3. A green ratio optimization system based on adaptive control according to the method of any one of claims 1 to 2, characterized in that, The system includes: The first module is used to convert the queue length of each phase into the queue length of the stage, and further into the stage adjustment time; The second module is used to redistribute the stage adjustment time to each stage according to the time constraints, and obtain the stage time for each stage. The third module is used to send the generated stage time to the signal machine for execution.

4. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 2.

Images