An adaptive signal coordination control method for shifted left-turn intersections
Through the adaptive signal coordination control method, traffic flow prediction and optimization algorithm are used to optimize signal timing, the problem of vehicle queue overflow in shift left-turn intersection signal control is solved, and the traffic signal control efficiency and driving efficiency are improved.
Patent Information
- Application Number
- CN202210643095.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-08
- Publication Date
- 2025-01-10
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, the signal control of the shifted left-turn intersection mainly adopts a fixed signal cycle method, which cannot effectively adapt to the dynamic fluctuations in the traffic flow at the intersection, resulting in the vehicle turning left-turned at the main intersection that may queuing overflow.
Adaptive signal coordination control method is adopted, by setting the signal phase sequence of the main intersection and the secondary intersection, initializing the signal period and green light duration, using traffic flow detection data and prediction models to perform short-term flow prediction, optimize signal timing to minimize Vehicle average delays are combined with genetic algorithms and particle swarm optimization algorithms to determine the optimal signal timing.
It improves the efficiency of traffic signal control, reduces traffic congestion, improves driving efficiency, and avoids the situation of overflowing vehicles turning left at the main intersection.
Smart Images
Figure CN117238149B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to traffic signal control technology, and specifically to an adaptive signal coordination control method for a shifted left-turn intersection. Background Art
[0002] As an intersection where multiple sources of traffic flow in different directions intersect, there are many conflict points, especially the conflict of left-turning traffic, which is one of the main reasons for the increase of intersection delays and even frequent congestion. The shifted left-turn intersection is an unconventional intersection design, which allows left-turning vehicles to turn to the shifted left-turn storage lane outside the opposite exit lane at the secondary intersection, eliminating the conflict between left-turning vehicles and straight-moving vehicles at the main intersection, and can complete the simultaneous release of left-turning and straight-moving vehicles in the same phase, reducing the number of phases in the cycle, improving the utilization rate of intersection time and space resources, and effectively reducing intersection delays. Reasonable setting of the shifted left-turn intersection can effectively alleviate the congestion of the intersection, thereby improving the travel comfort of citizens.
[0003] However, current research on signal control for displaced left-turn intersections mainly focuses on the method of fixed signal cycle control, that is, the intersection operates according to a pre-set control scheme, but the traffic flow at the intersection often has a certain dynamic volatility, and the timing control cannot fully adapt to and meet the changes in short-term traffic demand, which may lead to overflow of left-turn vehicles at the main intersection. Summary of the invention
[0004] The present application provides an adaptive signal coordination control method for a shifted left-turn intersection, which is used to solve the problem of left-turn vehicle queue overflow at a main intersection in the prior art.
[0005] The application provides an adaptive coordinated signal control method for a shifted left-turn intersection, the method comprising:
[0006] Step 1: according to the selected intersection, set the signal phase sequence of the main intersection and the secondary intersection; the main intersection is a shifted left-turn intersection, and the secondary intersection is an intersection in the east and west directions of the main intersection;
[0007] Step 2: Initialize the signal cycle of the intersection and the green light duration of each phase according to the historical traffic flow of the intersection;
[0008] Step 3: Obtain the flow arrival status of each entrance of the intersection based on the detection data collected by the detectors arranged at the intersection, and perform short-term flow prediction of each entrance through the prediction model;
[0009] Step 4: Taking the minimum average vehicle delay as the optimization goal, the coordination relationship of the main and pre-signals and the minimum green light time as the constraints, determine the signal timing optimization model;
[0010] Step 5: Solve the signal timing optimization model to obtain the optimal signal timing for the next signal cycle.
[0011] Optionally, according to the situation of the selected intersection, the signal phase sequence of the main intersection and the secondary intersection is set, including:
[0012] The phases include the east-west straight-forward left turn phase, the north-south straight-forward phase, and the north-south left turn phase; the phase sequence is north-south straight-forward - east-west straight-left - north-south left turn.
[0013] Optionally, according to the situation of the selected intersection, setting the signal phase sequence of the main intersection and the secondary intersection also includes:
[0014] The phases include the east-west straight left phase, the south straight left phase and the north straight left phase; the phase sequence is east-west straight left turn - south straight left - north straight left.
[0015] Optionally, the calculation formula for initializing the green light time of each phase is:
[0016]
[0017] L=∑ K (L s +IA) K (2)
[0018] Where L is the total signal loss time; Y is the total flow ratio; y i is the flow ratio of the i-th signal phase, L s The starting loss time is 3s for the yellow light time A.
[0019] Optionally, the flow arrival situation of each entrance of the intersection is obtained according to the detection data collected by the detectors arranged at the intersection, and the short-term flow prediction of each entrance is performed through the prediction model, including:
[0020] The traffic flow data of the detectors at each entrance and exit of the intersection are used to construct a sample set for the regression problem, and the data in the sample set is repaired, denoised and normalized.
[0021] Assume that the traffic volume in the current time period t is q i′ (t), the traffic flow in the next period is q i′ (t+1), then the corresponding training sample input is x i′ =(q i′ (t),q i′ (t-1),…,q i′ (t-n+1));
[0022] Set initial parameters, analyze the training set based on the genetic algorithm and select a suitable kernel function and find the penalty coefficient C and kernel parameter g with the best model performance; the initial parameters include the parameter range of the genetic algorithm, the population size, the number of iterations, and the crossover and mutation probabilities;
[0023] Finally, establish a prediction function and output the prediction results. The prediction function is as follows;
[0024]
[0025] In the formula, x i' Refers to the i'th sample among M sample data; α i' * , α i' is the support vector regression model parameter, and function K is the kernel function;
[0026] Predict the traffic volume in the next time period based on the given sample input.
[0027] Optionally, the average vehicle delay is determined by:
[0028]
[0029] Among them, the vehicles were delayed:
[0030]
[0031] is the average delay of vehicles in each phase i; q i is the traffic volume in phase i; C is the signal cycle length; λ is the green signal ratio of the lane group; X is the saturation of the lane group; c is the traffic capacity of the lane group; T is the analysis duration;
[0032] Constraints:
[0033] Constraint 1: To allow east-west left-turning vehicles to pass smoothly through the main intersection during the east-west straight left-turn phase, the green light duration of the pre-signal should not be greater than the green light duration of the east-west straight left-turn phase g y ≤g ew ;
[0034] Constraint 2: Because the east-west straight-moving vehicles conflict with the left-turning vehicles at the secondary intersection, the pre-signal green light closing time is the time when the east-west straight-moving vehicles arrive at the secondary intersection. Where L1 is the vertical distance between the east-west stop lines of the main intersection, V1 is the speed of the vehicle at the main intersection, L2 is the length of the shifted left-turn lane, and V2 is the speed of the vehicle between the main intersection and the secondary intersection;
[0035] Constraint 3: Because left-turning vehicles conflict with east-west through-going vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal duration should meet Where L3 is the length of the shifted left-turning vehicle crossing section, and V3 is the speed of the vehicle at the secondary intersection;
[0036] Constraint 4: Because left-turning vehicles conflict with south-north left-turning vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal green light should be turned on when the last south-north left-turning vehicle passes the secondary intersection.
[0037] Constraint 5: The minimum green light time should meet the needs of pedestrians crossing the street Where, L p is the length of pedestrian crossing, V p The pedestrian crossing speed is usually 1m / s, and I is the green light interval.
[0038] Optionally, the signal timing optimization model is solved to obtain the optimal signal timing for the next signal cycle in the following specific method:
[0039] Step 5.1, initialize the particle swarm size, number of iterations, inertia weight and learning factor parameters;
[0040] Step 5.2: Initialize the particle position and particle velocity and calculate the fitness value of the initialized particle to obtain the historical optimal position P of the particle. best and the global optimal position G of the group best ;
[0041] Step 5.3, update the velocity and position of the particle and calculate the fitness value of the updated particle;
[0042] Step 5.4, update the particle's historical optimal position Pbest and the group's global optimal position Gbest;
[0043] Step 5.5: Determine whether the particle swarm algorithm meets the termination condition. If the termination condition is met, output the result. Otherwise, repeat steps 5.3, 5.4 and 5.5 until the termination condition is met.
[0044] The present application improves the accuracy of traffic flow prediction by using genetic algorithms to seek optimal parameters in the traffic prediction model, and designs an optimization algorithm to solve the vehicle average delay index optimization model in real time. Compared with fixed signal timing, the adaptive coordinated signal control designed in the present application improves the efficiency of traffic signal control, thereby improving driving efficiency and reducing traffic congestion.
[0045] The present invention is described in further detail below in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 A schematic diagram of a shifted left-turn intersection provided in an embodiment of the present application;
[0047] Figure 2 A signal phase sequence diagram of a shifted left-turn intersection provided in an embodiment of the present application;
[0048] Figure 3 A signal timing diagram for a shifted left-turn intersection provided in an embodiment of the present application;
[0049] Figure 4 A schematic flow chart of an adaptive signal coordination control method for a shifted left-turn intersection provided in an embodiment of the present application. DETAILED DESCRIPTION
[0050] A coordinated signal control method for a shifted left-turn intersection, comprising the following specific steps:
[0051] Step 1: According to the specific conditions of the selected intersection, reasonably set the signal phase sequence of the main intersection and the secondary intersection.
[0052] The signal phase sequence of primary and secondary intersections can be reasonably determined based on the specific geometric layout of the intersection and the actual traffic flow characteristics.
[0053] A feasible phase includes an east-west straight-running left turn phase, a north-south straight-running phase, and a north-south left turn phase; the phase sequence is north-south straight-running (pre-signal left turn) - east-west straight-running left - north-south left turn.
[0054] Another feasible phase includes the east-west straight left phase, the south straight left phase and the north straight left phase; the phase sequence is east-west straight left turn - south straight left (east warning signal left turn) - north straight left (west warning signal left turn).
[0055] This description takes the east-west direction of the intersection as a shifted left turn mode, and the north-south direction as a conventional intersection as a specific embodiment. The main signal phase 1 is for north-south straight driving. When the north-south straight driving green light is on, the traffic signal of the main intersection in the east-west direction needs to be coordinated with the pre-signal to allow the east-west left-turning traffic to enter the shifted left-turn storage lane to wait for the east-west left turn. Phase 2 is for east-west straight driving to turn left, and phase 3 is for north-south left turning. The phase sequence and signal timing design scheme of this article is as follows Figure 1 , Figure 2 shown.
[0056] Step 2: Initialize the signal cycle of the intersection and the green light duration of each phase according to the historical traffic flow of the intersection.
[0057] Specifically, the initialization green light duration of each phase of the intersection is determined by the Webster method;
[0058] The calculation formula for the effective green light duration of the i-th signal phase is:
[0059]
[0060] L=∑ K (L s +IA) K (2)
[0061] Where L is the total signal loss time; Y is the total flow ratio; y i is the flow ratio of the i-th signal phase, usually L s The starting loss time is 3s for the yellow light time A.
[0062] Step 3: Obtain the flow arrival status of each entrance of the intersection based on the detection data collected by the detectors arranged at the intersection, and perform short-term flow prediction of each entrance through the prediction model.
[0063] This application uses support vector regression to establish a short-term traffic flow prediction model. First, the historical traffic flow data of the detector at each entrance and exit of the intersection is repaired, denoised, and normalized to improve the accuracy of the prediction model.
[0064] Assume that the traffic volume in the current time period t is q i' (t), the traffic flow in the next period is q i' (t+1), then the corresponding training sample input is x i' =(q i' (t),q i' (t-1),…,q i' (t-n+1)).
[0065] Secondly, the initial parameters of the genetic algorithm, such as parameter range, population size, number of iterations, and crossover and mutation probabilities, are set. The genetic algorithm is used to analyze the training set and select the appropriate kernel function and find the penalty coefficient C and kernel parameter g for optimal model performance.
[0066] Finally, the prediction function is established and the prediction result is output. The prediction function is shown in formula (3).
[0067]
[0068] In the formula, x i' Refers to the i'th sample among M sample data; α i′ * , α i′ is the support vector regression model parameter, function K is the kernel function, and this application uses the Gaussian radial basis kernel function:
[0069] At the same time, the traffic volume in the next time period is predicted based on the given sample input.
[0070] Step 4: Taking the minimum average vehicle delay as the optimization goal, the coordination relationship of the main and pre-signals and the minimum green light time as the constraints, determine the signal timing optimization model;
[0071] Optimization goal:
[0072] Among them, the vehicles were delayed:
[0073]
[0074] In the formula, is the average delay of vehicles in each phase i; q i is the traffic flow in phase i; C is the signal cycle length; λ is the green-signal ratio of the lane group; X is the saturation of the lane group; c is the traffic capacity of the lane group; T is the analysis duration.
[0075] Constraints:
[0076] Constraint 1: In order for the east-west left-turning vehicles to pass through the main intersection smoothly during the east-west straight left-turn phase, i.e., phase 2, the green light duration of the pre-signal should not be greater than the green light duration of phase 2 g y ≤g2;
[0077] Constraint 2: Because the east-west straight-moving vehicles conflict with the left-turning vehicles at the secondary intersection, the pre-signal green light closing time is the time when the east-west straight-moving vehicles arrive at the secondary intersection. Where L1 is the vertical distance between the east-west stop lines of the main intersection, V1 is the speed of the vehicle at the main intersection, L2 is the length of the shifted left-turn lane, and V2 is the speed of the vehicle between the main intersection and the secondary intersection;
[0078] Constraint 3: Because left-turning vehicles conflict with east-west through-going vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal duration should meet Where L3 is the length of the shifted left-turning vehicle crossing section, and V3 is the speed of the vehicle at the secondary intersection;
[0079] Constraint 4: Because left-turning vehicles conflict with south-north left-turning vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal green light should be turned on when the last south-north left-turning vehicle passes the secondary intersection.
[0080] Constraint 5: The minimum green light time should meet the needs of pedestrians crossing the street Where, L p is the length of pedestrian crossing, V pThe pedestrian crossing speed is usually 1m / s, and I is the green light interval.
[0081] Step 5: Solve the signal timing optimization model to obtain the optimal signal timing for the next signal cycle;
[0082] This application uses a particle swarm optimization algorithm to solve the signal timing optimization model.
[0083] Step 5.1, initialize the parameters such as particle swarm size, number of iterations, inertia weight and learning factor.
[0084] Step 5.1, initialize particle position and velocity.
[0085] Step 5.2: Initialize the particle position and particle velocity and calculate the fitness value of the initialized particle to obtain the historical optimal position P of the particle. best and the global optimal position G of the group best .
[0086] Step 5.3: Update the velocity and position of the particle and calculate the fitness value of the updated particle.
[0087] Specifically, according to the particle's historical optimal position P best and the global optimal position G of the group best Update the speed and position of the particle and calculate the fitness value of the updated particle. The speed and position update formulas are:
[0088]
[0089] Where i is the particle number, d is the particle dimension number, k is the number of iterations, which is 100 in this application, ω is the inertia weight, which is 1.2, c1 and c2 are the individual learning factor and the group learning factor, with the value c1=c2=2, and r1 and r2 are random numbers between [0,1].
[0090] Step 5.4: Update the particle's historical optimal position P best and the global optimal position G of the group best .
[0091] Step 5.5: Determine whether the algorithm meets the termination condition. If so, output the result. Otherwise, repeat steps 5.3, 5.4 and 5.5 until the termination condition is met.
[0092] For ease of expression, this application only explains the situation of setting a shifted left-turn intersection in the north-south direction. This application needs to change the corresponding parameter conditions when applied to the situation of setting a shifted left-turn intersection in other directions. At the same time, the above-mentioned formula values of this application are only for detailed explanation of the calculation process of the model, and the values of the corresponding parameters can be changed according to the specific situation.
[0093] The present application improves the accuracy of traffic flow prediction by using genetic algorithms to seek optimal parameters in the traffic prediction model, and designs an optimization algorithm to solve the vehicle average delay index optimization model in real time. Compared with fixed signal timing, the adaptive coordinated signal control designed in the present application improves the efficiency of traffic signal control, thereby improving driving efficiency and reducing traffic congestion.
Claims
1. An adaptive coordinated signal control method for a shifted left-turn intersection, characterized in that: The method comprises: Step 1: according to the selected intersection, set the signal phase sequence of the main intersection and the secondary intersection; the main intersection is a shifted left-turn intersection, and the secondary intersection is an intersection in the east and west directions of the main intersection; Step 2: Initialize the signal cycle of the intersection and the green light duration of each phase according to the historical traffic flow of the intersection; Step 3: Obtain the flow arrival status of each entrance of the intersection based on the detection data collected by the detectors arranged at the intersection, and perform short-term flow prediction of each entrance through the prediction model; Step 4: Taking the minimum average vehicle delay as the optimization goal, the coordination relationship of the main and pre-signals and the minimum green light time as the constraints, determine the signal timing optimization model; Step 5: Solve the signal timing optimization model to obtain the optimal signal timing for the next signal cycle; The average vehicle delay is determined by the following method: Among them, the vehicles were delayed: is the average delay of vehicles in each phase i; q i is the traffic flow in phase i; C is the signal cycle length; λ is the green signal ratio of the lane group; X i is the lane group saturation at phase i; c is the lane group capacity; T is the analysis duration; Constraints: Constraint 1: To allow east-west left-turning vehicles to pass smoothly through the main intersection during the east-west straight left-turn phase, the green light duration of the pre-signal should not be greater than the green light duration of the east-west straight left-turn phase g y ≤g ew ; Constraint 2: Because the east-west straight-moving vehicles conflict with the left-turning vehicles at the secondary intersection, the pre-signal green light closing time is the time when the east-west straight-moving vehicles arrive at the secondary intersection. Where L1 is the vertical distance between the east-west stop lines of the main intersection, V1 is the speed of the vehicle at the main intersection, L2 is the length of the shifted left-turn lane, and V2 is the speed of the vehicle between the main intersection and the secondary intersection; Constraint 3: Because left-turning vehicles conflict with east-west through-going vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal duration should meet Where L3 is the length of the shifted left-turning vehicle crossing section, and V3 is the speed of the vehicle at the secondary intersection; Constraint 4: Because left-turning vehicles conflict with south-north left-turning vehicles when entering the left-turning vehicle storage lane at the secondary intersection, the pre-signal green light should be turned on when the last south-north left-turning vehicle passes the secondary intersection. Constraint 5: The minimum green light time should meet the needs of pedestrians crossing the street Where, L p is the length of pedestrian crossing, V p The pedestrian crossing speed is usually 1m / s, and I is the green light interval.
2. The method according to claim 1, characterized in that According to the selected intersection, set the signal phase sequence of the main intersection and the secondary intersection, including: The phases include the east-west straight-forward left turn phase, the north-south straight-forward phase, and the north-south left turn phase; the phase sequence is north-south straight-forward - east-west straight-left - north-south left turn.
3. The method according to claim 1, characterized in that According to the selected intersection, the signal phase sequence of the main intersection and the secondary intersection is set, which also includes: The phases include the east-west straight left phase, the south straight left phase and the north straight left phase; the phase sequence is east-west straight left turn - south straight left - north straight left.
4. The method according to claim 1, characterized in that: The calculation formula for initializing the green light time of each phase is: L=∑ K (L s +I-A) K (2) Where L is the total signal loss time; Y is the total flow ratio; y i is the flow ratio of the i-th signal phase, L s The starting loss time is 3s for the yellow light time A.
5. The method according to claim 1, characterized in that The traffic arrival situation at each entrance of the intersection is obtained based on the detection data collected by the detectors arranged at the intersection, and the short-term traffic flow prediction of each entrance is carried out through the prediction model, including: The traffic flow data of the detectors at each entrance and exit of the intersection are used to construct a sample set for the regression problem, and the data in the sample set is repaired, denoised and normalized. Assume that the traffic volume in the current time period t is q i' (t), the traffic flow in the next period is q i' (t+1), then the corresponding training sample input is x i' =(q i' (t),q i' (t-1),…,q i' (t-n+1)); Set initial parameters, analyze the training set based on the genetic algorithm and select a suitable kernel function and find the penalty coefficient C and kernel parameter g with the best model performance; the initial parameters include the parameter range of the genetic algorithm, the population size, the number of iterations, and the crossover and mutation probabilities; Finally, establish a prediction function and output the prediction results. The prediction function is as follows; In the formula, x i' Refers to the i'th sample among M sample data; α i' is the support vector regression model parameter, and function K is the kernel function; Predict the traffic volume in the next time period based on the given sample input.
6. The method according to claim 1, characterized in that The specific method for solving the signal timing optimization model and obtaining the optimal signal timing for the next signal cycle is: Step 5.1, initialize the particle swarm size, number of iterations, inertia weight and learning factor parameters; Step 5.2: Initialize the particle position and particle velocity and calculate the fitness value of the initialized particle to obtain the historical optimal position P of the particle. best and the global optimal position G of the group best ; Step 5.3, update the velocity and position of the particle and calculate the fitness value of the updated particle; Step 5.4, update the particle's historical optimal position Pbest and the group's global optimal position Gbest; Step 5.5: Determine whether the particle swarm algorithm meets the termination condition. If the termination condition is met, output the result. Otherwise, repeat steps 5.3, 5.4 and 5.5 until the termination condition is met.