A method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithm and a computer storage medium.
By introducing dedicated lanes and phases for autonomous vehicles at intersections, and combining this with genetic algorithms to optimize lane organization and signal control, the scheduling problem of autonomous and manually driven vehicles in mixed traffic environments has been solved, resulting in more efficient intersection passage and improved adaptability and management efficiency of the traffic system.
Patent Information
- Application Number
- CN202510070765.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-16
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2045-01-16
AI Technical Summary
In mixed traffic environments, the unified control of intersection scheduling strategies for autonomous vehicles and human-driven vehicles leads to mutual interference, affecting the driving advantages of autonomous vehicles. Existing technologies lack dynamic adaptability and dedicated lanes and phases.
By introducing dedicated lanes and phases for autonomous vehicles at intersections, a collaborative optimization model for lane organization and signal control based on genetic algorithms is constructed. By constructing objective functions, setting constraints, and solving the problem using genetic algorithms, the average vehicle delay is optimized, and a signal timing scheme is designed to achieve separate scheduling of autonomous and manually driven vehicles.
It significantly improves the traffic efficiency at intersections, reduces average vehicle delays, enhances the adaptability and decision-making time of intelligent transportation systems, and provides better traffic management solutions.
Smart Images

Figure CN119889037B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of traffic control technology, and in particular relates to a method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, as well as a computer storage medium. Background Technology
[0002] Currently, autonomous driving technology is considered a crucial trend in future transportation development. However, in the short term, autonomous vehicles are unlikely to completely replace manually driven vehicles. Therefore, in mixed traffic environments, how to accurately and effectively manage autonomous vehicles (CAVs) and manually driven vehicles (HVs) and maximize control efficiency has become a pressing issue that needs to be addressed.
[0003] In current research, most scheduling strategies for mixed traffic flow at intersections use a uniform control method, where autonomous vehicles and human-driven vehicles use the same intersection approach lanes and signal phases. This leads to mutual interference between autonomous and human-driven vehicles, greatly affecting the ability of autonomous vehicles to leverage their driving advantages.
[0004] To address the aforementioned issues, this invention proposes introducing dedicated lanes and phases for autonomous vehicles at intersections and establishes a collaborative optimization model for lane organization and signal control at intersections considering mixed traffic flows. This separates autonomous vehicles from manually driven vehicles in both time and space, allowing autonomous vehicles to fully utilize their driving characteristics and effectively reducing average vehicle delays at intersections. Summary of the Invention
[0005] The purpose of this invention is to provide a method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, as well as a computer storage medium, to solve the problems of current scheduling strategies for mixed traffic flow at intersections, such as unified control, mutual crossing, and the lack of dedicated lanes and phases for autonomous driving.
[0006] To achieve the above objectives, this invention provides a method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, comprising the following steps:
[0007] Step 1: Construct an objective function with average vehicle delay as the objective variable;
[0008] Step 2: Based on traffic flow theory, set constraints and construct a collaborative optimization model for lane organization and signal control at mixed-traffic intersections; the constraints include lane layout constraints, traffic flow allocation constraints, and signal timing constraints.
[0009] Step 3: Solve the objective function based on the genetic algorithm to obtain the optimal solution; initialize the model parameters, generate an initial population, design the signal timing scheme using the Webster method, define the fitness function, perform genetic operations on the samples, and iterate continuously to obtain the scheme with the minimum average delay per vehicle as the optimal solution.
[0010] Preferably, in step 1, the objective function is constructed with average vehicle delay as the objective variable as follows:
[0011] Z = mind;
[0012]
[0013] In the formula, d represents the average delay per vehicle, Z is the set of vehicle types, Z = {A, H}, where A and H represent autonomous vehicles and manually driven vehicles, respectively, and Q... i,j For the traffic volume of the Class J functional lane at the intersection i approach, d i,j This indicates the average vehicle delay in lane j of the i-direction approach, Y. i,j Let β be the saturation level of lane j in the direction of intersection i, T be the signal cycle duration, and β be the signal saturation level. i,j The green light ratio is the green light ratio of the Class J functional lane at the i-th approach of the intersection.
[0014] Preferably, the lane layout constraints are as follows:
[0015] The layout of dedicated entrance lanes for autonomous driving follows the following formula:
[0016] M(sgn(f i,1 )-1)≤g i,1 ≤Msgn(f i,1 )
[0017] Where M is a sufficiently large positive number; sgn() is the sign function, which takes the value 1 when the parameter inside the function is positive, and 0 otherwise; f i,1 The number of lanes for the dedicated autonomous driving lanes at the intersection i-entry direction; g i,1 The green light time for the dedicated lane for autonomous driving at the intersection's i-entry direction;
[0018] When the number of dedicated autonomous driving approach lanes is not zero, the dedicated autonomous driving phase at the intersection is activated, at which point f i,1 >0, sgn(f i,1 ) = 1; When no dedicated autonomous driving approach lane is provided, the dedicated autonomous driving phase at the intersection is closed, and f = 1. i,1 =0, sgn(f i,1 ) = 0;
[0019] For manual driving lanes, certain types of lanes cannot coexist in the same direction of entry, and the constraints follow the formula below:
[0020]
[0021] Among them, f i,2 ,f i,4 ,f i,5 ,f i,6 ,f i,7 ,f i,8 These refer to the number of lanes at the intersection i-entry direction: left-turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane.
[0022] The number of lane groups is guaranteed to be a non-negative integer, constrained by the following formula:
[0023]
[0024] In the formula, f i,j Indicates the number of lanes of type j in the direction of intersection i;
[0025] For each direction of entry, the number of dedicated lanes for autonomous driving is equal to the number of dedicated lanes for autonomous driving deployed in that direction, and the number of regular lanes is equal to the number of entry lanes in that direction minus the number of dedicated lanes for autonomous driving. The number of lanes is defined by the following formula:
[0026]
[0027] in, Fi represents the number of dedicated lanes for automated driving and the number of lanes for manual driving at intersection i, respectively. Fi is the number of lanes at intersection i; i represents the direction of the approach lane. In a cross intersection, i∈{1,2,3,4}, representing the north approach lane, south approach lane, west approach lane, and east approach lane, respectively.
[0028] Preferably, the traffic flow allocation constraints include traffic flow constraints for dedicated autonomous driving lanes, traffic volume constraints for manual driving lanes, turning traffic flow balance constraints, multi-turn lane traffic flow balance constraints, and traffic flow allocation equilibrium constraints, with specific expressions as follows:
[0029] When a dedicated lane for autonomous driving exists, all autonomous vehicles use that lane; otherwise, the traffic flow is zero. The traffic flow constraint Q for the dedicated autonomous driving lane is... i,1 as follows:
[0030]
[0031] In the formula, k is the flow direction symbol, k∈{L,T,R}, which represent left turn, straight ahead, and right turn, respectively; Traffic flow of autonomous vehicles in direction k of the i-import lane;
[0032] Traffic volume constraints for manual driving lanes are as follows:
[0033]
[0034] Among them, Q i,k Traffic volume using the manual driving lane in the direction of i entrance k; Traffic flow of manually driven vehicles in direction k of the i-entry lane;
[0035] To ensure that all turning traffic flows at each approach lane are fully allocated to their respective lanes, the following formula is used to achieve traffic flow balance constraints at each turn:
[0036] Q i,2 f i,2 +Q i,5 f i,5 +Q i,7,L f i,7 +Q i,8,L f i,8 =Q i,L ;
[0037] Q i,3 f i,3 +Q i,6,R f i,6 +Q i,7,R f i,7 +Q i,8,R f i,8 =Q i,R ;
[0038] Q i,4 f i,4 +Q i,5,T f i,5 +Q i,6,T f i,6 +Q i,8,T f i,8 =Q i,T ;
[0039] Among them, Q i,j,k Let Q be the traffic flow of lane j in the i-th entrance lane towards lane k. i,k Let Q be the traffic volume of vehicles traveling in the manual driving lane at entrance i in direction k, k∈{L,T,R}. i,j Let i represent the traffic volume of lane j with function at intersection i, where i∈{1,2,3,4} represents the north entrance lane, south entrance lane, west entrance lane, and east entrance lane, respectively; j represents the lane function, where j∈{1,2,...,6,7,8} represents the dedicated lane for automated driving, left turn lane, right turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane, respectively.
[0040] To ensure that various types of entry lanes are correctly matched with the corresponding turning traffic flow, the following formula is used to achieve traffic flow balance constraints for multiple turning lanes:
[0041] Q i,5,L +Q i,5,T =Q i,5 ;
[0042] Q i,6,R +Q i,6,T =Q i,6 ;
[0043] Q i,7,L +Q i,7,R =Q i,7 ;
[0044] Q i,8,T +Q i,8,L +Q i,8,R =Q i,8 ;
[0045] When the i-th approach lane has multiple lane groups facing a certain traffic flow, the traffic flow in that direction is distributed among the lane groups according to the principle of equal saturation. The traffic flow distribution balance constraint is achieved by the following formula:
[0046]
[0047] M(2-sgn(f i,j1 )-sgn(f i,j2 ))≥Y i,j1 -Y i,j2 ≥-M(2-sgn(f i,j1 )-sgn(f i,j2 )),
[0048]
[0049] In the formula, ε is an infinitesimal quantity used to ensure the computability of the model when the lane group does not exist; δ is the vehicle start-up loss time; C i,j To ensure the traffic capacity of the Class J functional lane in the direction of intersection i, g i,j This indicates the green light signal time for lane j in the direction of import i, Y. i,j Indicates the saturation of lane j in the direction of import i, f i,j1 Indicates the number of lanes of type j1 in the direction of import i, f i,j2 Indicates the number of Class J2 lanes in the direction of import i, Y i,j1 Indicates the saturation of lane j1 in the direction of import i, Y i,j2 This indicates the saturation of lane type j2 in the direction of import i.
[0050] Preferably, the signal timing constraints are as follows: Considering the inter-autonomous vehicle weaving through traffic at the intersection, an additional dedicated autonomous driving phase is added to the existing signal phases. The design incorporates a "double-ring-three-group" structure for this dedicated autonomous driving phase. The autonomous driving traffic flows from each approach direction form the third dedicated autonomous driving phase. Specific constraints are as follows:
[0051] The green light duration in signal timing is constrained by the following formula:
[0052] g min ≤g i,j ≤g max ;
[0053] Among them, g i,j The green light time for lane j in the direction of intersection i; g min and g max These are the minimum and maximum green light durations for the signal phase, respectively.
[0054] The signal group needs to be determined in advance, and the signal flow to the green light time in the signal group should be equal, defined by the following formula:
[0055]
[0056] Among them, g m Ω represents the green light duration for the m-th signal group. m For the m-th signal group; g i,j(i,j) This indicates the green light signal time for lane type j(i,j) in the direction of entry i; a signal group refers to a combination of lane groups that can pass at the same time.
[0057] The signal cycle duration T at the intersection is no greater than the maximum cycle duration T. max Not less than the minimum period duration T min The expression is as follows:
[0058] T min ≤T≤T max ;
[0059]
[0060] In the formula, σ represents the intersection clearing time.
[0061] Preferably, the fitness function f(X) in step 3 is as follows:
[0062]
[0063] In the formula, d(X) represents the average delay per vehicle.
[0064] The present invention also provides a computer storage medium storing computer-readable instructions thereon, which, when executed by a computer, implement the above-described method.
[0065] Therefore, this invention employs a genetic algorithm-based method for the coordinated optimization of lane organization and signal control at mixed-traffic intersections, along with a computer storage medium. By integrating collected traffic data, performing information integration, feature analysis, and key parameter optimization based on a pre-defined algorithm framework, control decisions are made to improve intersection traffic efficiency. By combining the characteristics of autonomous vehicles and the dynamic behavior of manually driven vehicles, the invention addresses the lack of adaptability in existing traffic control strategies. Furthermore, compared to traditional signal control methods, it significantly improves optimization speed, constructs a coordinated optimization model for lane organization and signal control at mixed-traffic intersections, and achieves better traffic efficiency. This overcomes the shortcomings of existing intersection traffic flow control technologies, which require overcoming a single control strategy and lack dynamic adaptability. This is crucial for intelligent transportation systems to predict traffic congestion and provide real-time scheduling, while also providing traffic managers with more decision-making time.
[0066] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description
[0067] Figure 1 This is a flowchart of a method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, according to the present invention.
[0068] Figure 2 This is a diagram showing the intersection lane organization layout according to an embodiment of the present invention;
[0069] Figure 3 This is a graph showing the variation of average vehicle delay at intersections under different CAV penetration rates according to an embodiment of the present invention.
[0070] Figure 4 This is a graph showing the reduction in average vehicle delay (HV), average vehicle delay (CAV), and average vehicle delay at intersections under different CAV penetration rates according to an embodiment of the present invention. Detailed Implementation
[0071] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.
[0072] Please see Figures 1-2 A method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms includes the following steps:
[0073] Step 1: Construct the objective function with average vehicle delay as the objective variable; the specific expression is as follows:
[0074] Z = mind;
[0075]
[0076] In the formula, d represents the average delay per vehicle, Z is the set of vehicle types, Z = {A, H}, where A and H represent autonomous vehicles and manually driven vehicles, respectively, and Q... i,j For the traffic volume of the Class J functional lane at the intersection i approach, d i,j This indicates the average vehicle delay in lane j of the i-direction approach, Y. i,j Let β be the saturation level of lane j in the direction of intersection i, T be the signal cycle duration, and β be the signal saturation level. i,j The green light ratio is the green light ratio of the Class J functional lane at the i-th approach of the intersection.
[0077] Step 2: Based on traffic flow theory, set constraints and construct a collaborative optimization model for lane organization and signal control at mixed-traffic intersections; the constraints include lane layout constraints, traffic flow allocation constraints, and signal timing constraints.
[0078] The specific constraints on lane layout are as follows:
[0079] The layout of dedicated entrance lanes for autonomous driving follows the following formula:
[0080] M(sgn(f i,1 )-1)≤g i,1 ≤Msgn(f i,1 )
[0081] Where M is a sufficiently large positive number; sgn() is the sign function, which takes the value 1 when the parameter inside the function is positive, and 0 otherwise; f i,1 The number of lanes for the dedicated autonomous driving lanes at the intersection i-entry direction; g i,1 The green light time for the dedicated lane for autonomous driving at the intersection's i-entry direction;
[0082] When the number of dedicated autonomous driving approach lanes is not zero, the dedicated autonomous driving phase at the intersection is activated, at which point f i,1 >0, sgn(f i,1 ) = 1; When no dedicated autonomous driving approach lane is provided, the dedicated autonomous driving phase at the intersection is closed, and f = 1. i,1 =0, sgn(f i,1 ) = 0;
[0083] For manual driving lanes, certain types of lanes cannot coexist in the same direction of entry, and the constraints follow the formula below:
[0084]
[0085] Among them, f i,2 ,f i,4 ,f i,5 ,f i,6 ,f i,7 ,f i,8 These refer to the number of lanes at the intersection i-entry direction: left-turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane.
[0086] The number of lane groups is guaranteed to be a non-negative integer, constrained by the following formula:
[0087]
[0088] In the formula, f i, j represents the number of lanes of type j in the direction of intersection i's approach;
[0089] For each direction of entry, the number of dedicated lanes for autonomous driving is equal to the number of dedicated lanes for autonomous driving deployed in that direction, and the number of regular lanes is equal to the number of entry lanes in that direction minus the number of dedicated lanes for autonomous driving. The number of lanes is defined by the following formula:
[0090]
[0091] in, Fi represents the number of dedicated lanes for automated driving and the number of lanes for manual driving at intersection i, respectively. Fi is the number of lanes at intersection i; i represents the direction of the approach lane. In a cross intersection, i∈{1,2,3,4}, representing the north approach lane, south approach lane, west approach lane, and east approach lane, respectively.
[0092] Traffic flow allocation constraints include traffic flow constraints for dedicated lanes for autonomous driving, traffic volume constraints for lanes for manual driving, traffic flow balance constraints for turning lanes, traffic flow balance constraints for multi-turn lanes, and traffic flow distribution equilibrium constraints. The specific expressions are as follows:
[0093] When a dedicated lane for autonomous driving exists, all autonomous vehicles use that lane; otherwise, the traffic flow is zero. The traffic flow constraint Q for the dedicated autonomous driving lane is... i,1 as follows:
[0094]
[0095] In the formula, k is the flow direction symbol, k∈{L,T,R}, which represent left turn, straight ahead, and right turn, respectively; Traffic flow of autonomous vehicles in direction k of the i-import lane;
[0096] Traffic volume constraints for manual driving lanes are as follows:
[0097]
[0098] Among them, Q i,k Traffic volume using the manual driving lane in the direction of i entrance k; Traffic flow of manually driven vehicles in direction k of the i-entry lane;
[0099] To ensure that all turning traffic flows at each approach lane are fully allocated to their respective lanes, the following formula is used to achieve traffic flow balance constraints at each turn:
[0100] Q i,2 f i,2 +Q i,5 f i,5 +Q i,7,L f i,7 +Q i,8,L f i,8 =Q i,L ;
[0101] Q i,3 f i,3 +Q i,6,R f i,6 +Q i,7,R f i,7 +Q i,8,R f i,8 =Q i,R ;
[0102] Q i,4 f i,4 +Q i,5,T f i,5 +Q i,6,T f i,6 +Q i,8,T f i,8 =Q i,T ;
[0103] Among them, Q i,j,k Let Q be the traffic flow of lane j in the i-th entrance lane towards lane k. i,k Let Q be the traffic volume of vehicles traveling in the manual driving lane at entrance i in direction k, k∈{L,T,R}. i,j Let i represent the traffic volume of lane j with function at intersection i, where i∈{1,2,3,4} represents the north entrance lane, south entrance lane, west entrance lane, and east entrance lane, respectively; j represents the lane function, where j∈{1,2,...,6,7,8} represents the dedicated lane for automated driving, left turn lane, right turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane, respectively.
[0104] To ensure that various types of entry lanes are correctly matched with the corresponding turning traffic flow, the following formula is used to achieve traffic flow balance constraints for multiple turning lanes:
[0105] Qi,5,L +Q i,5,T =Q i,5 ;
[0106] Q i,6,R +Q i,6,T =Q i,6 ;
[0107] Q i,7,L +Q i,7,R =Q i,7 ;
[0108] Q i,8,T +Q i,8,L +Q i,8,R =Q i,8 ;
[0109] When there are multiple lane groups facing a certain traffic flow at the i-th approach, the traffic flow in that direction is distributed among the lane groups according to the principle of equal saturation. The traffic flow distribution balance constraint is achieved by the following formula:
[0110]
[0111] M(2-sgn(f i,j1 )-sgn(f i,j2 ))≥Y i,j1 -Y i,j2 ≥-M(2-sgn(f i,j1 )-sgn(f i,j2 )),
[0112]
[0113] In the formula, ε is an infinitesimal quantity used to ensure the computability of the model when the lane group does not exist; δ is the vehicle start-up loss time; C i,j To ensure the traffic capacity of the Class J functional lane in the direction of intersection i, g i,j This indicates the green light signal time for lane j in the direction of import i, Y. i,j Indicates the saturation of lane j in the direction of import i, f i,j1 Indicates the number of lanes of type j1 in the direction of import i, f i,j2 Indicates the number of Class J2 lanes in the direction of import i, Y i,j1 Indicates the saturation of lane j1 in the direction of import i, Y i,j2 This indicates the saturation of lane type j2 in the direction of import i.
[0114] The signal timing constraints are as follows: Considering the weaving passage of autonomous vehicles, an additional dedicated autonomous driving phase is added to the existing signal phases. The design incorporates a "double-ring-three-group" structure for this dedicated autonomous driving phase. The autonomous driving traffic flows from each approach direction form the third dedicated autonomous driving phase. Specific constraints are as follows:
[0115] The green light duration in signal timing is constrained by the following formula:
[0116] g min ≤g i,j ≤g max ;
[0117] Among them, g i,j The green light time for lane j in the direction of intersection i; g min and g max These are the minimum and maximum green light durations for the signal phase, respectively.
[0118] The signal group needs to be determined in advance, and the signal flow to the green light time in the signal group should be equal, defined by the following formula:
[0119]
[0120] Among them, g m Ω represents the green light duration for the m-th signal group. m For the m-th signal group; g i,j(i,j) This indicates the green light signal time for lane type j(i,j) in the direction of entry i; a signal group refers to a combination of lane groups that can pass at the same time.
[0121] The signal cycle duration T at the intersection is no greater than the maximum cycle duration T. max Not less than the minimum period duration T min The expression is as follows:
[0122] T min ≤T≤T max ;
[0123]
[0124] In the formula, σ represents the intersection clearing time.
[0125] Step 3: Solve the objective function using a genetic algorithm to obtain the optimal solution; initialize the model parameters, generate an initial population, design a signal timing scheme using the Webster method, define the fitness function, perform genetic operations on the samples, and iterate to obtain the scheme with the minimum average vehicle delay as the optimal solution; the specific process is as follows:
[0126] S1. Initialize parameters:
[0127] Key parameters for initializing the model include lane combination type, traffic flow, autonomous vehicle penetration rate, lane group saturation flow rate, analysis time length, population size, crossover and mutation probabilities, and the maximum number of algorithm iterations. These parameters are set based on prior research and field data analysis to ensure the model performs well in real-world applications.
[0128] S2. Generate the initial population:
[0129] Based on the set population size, a series of possible schemes are randomly generated. These schemes include all possible combinations of different CAV penetration rates and different lane organization schemes, ensuring the diversity of the initial population and providing a broad gene pool for the operation of the genetic algorithm.
[0130] S3. Timing Scheme Design:
[0131] For each lane organization scheme generated in step S2, a signal timing scheme is designed using the Webster method. The Webster method is a classic signal timing design method that can effectively adjust the green light time of the signal according to traffic flow demand, thereby optimizing the traffic flow efficiency.
[0132] S4. Fitness function setting:
[0133] A fitness function is defined to evaluate the effectiveness of each sample's solution. The fitness function is primarily based on the average vehicle delay time, ensuring that the algorithm prioritizes solutions that minimize delay. After the fitness of each sample is calculated, a corresponding selection probability is assigned to each sample based on its fitness value for selection in subsequent iterations. The fitness function is defined by the following formula:
[0134]
[0135] S5. Genetic manipulation and sample update:
[0136] In this step, samples are selected to enter the next generation based on the sampling probabilities calculated in the previous step. After selection, genetic operations, including crossover and mutation, are performed on the selected samples to generate a new population.
[0137] S6. Iteration and Convergence
[0138] Repeat steps S3 and S5 until the preset number of iterations is reached or the quality of the solution no longer shows significant improvement. At the end of the iteration process, the lane organization and signal control scheme that performs best under different autonomous vehicle penetration rates across all rounds, i.e., the scheme with the minimum average vehicle delay, is the optimal solution under that autonomous vehicle penetration rate.
[0139] This invention is based on Figure 2Taking the intersection lane organization shown in Figure 1 and the intersection traffic flow shown in Table 1 as examples, a case study is conducted to verify the effectiveness of the collaborative optimization model of intersection lane organization and signal control under mixed traffic conditions. The relevant parameter values during the modeling process are shown in Table 2.
[0140] Table 2 Initial Traffic Volume at the Intersection
[0141] flow straight Turn left Turn right East 811 268 132 West 968 339 155 South 1333 251 301 north 887 236 188
[0142] Table 3 Model parameter values
[0143]
[0144] Based on the above parameter settings, a designed genetic algorithm is used to solve the model. Initially, 200 lane organization schemes are randomly generated as the initial population. In each generation of evolution, the fitness value of each individual is calculated first, which is the reciprocal of its average vehicle delay. Then, selection, crossover, and mutation operations are performed according to a certain probability to generate a new generation of population. By running the algorithm independently multiple times, the optimal solution from the multiple runs is finally taken as the optimal lane organization and signal control scheme for the intersection.
[0145] This paper sets six typical CAV penetration rates, and the solution results for different penetration rates are shown in Tables 3 and 4. Table 3 shows the minimum average vehicle delay and the optimal number of CAV-dedicated lanes under different penetration rates. The classic model (intersection control model without dedicated autonomous driving approach lanes and dedicated phases) yields an average vehicle delay of 60.577s. When the CAV penetration rate is 0-42%, no dedicated CAV lanes are set up, and the average vehicle delay at the intersection is equal to that of the classic model. When the CAV penetration rate is 42-73%, one dedicated CAV lane is set up. When the CAV penetration rate increases to 73% or above, two dedicated CAV lanes are set up.
[0146] Table 3 Minimum Average Vehicle Delay under Different CAV Penetration Rates
[0147]
[0148]
[0149] Table 4 shows the signal timing schemes for intersections under different CAV penetration rates. When the CAV penetration rate is 0-42%, no dedicated CAV phase is set, and the signal timing scheme for the intersection is the same as when the CAV penetration rate is 0%. When the CAV penetration rate is higher than 42%, a dedicated CAV phase is set. As the CAV penetration rate increases, the signal cycle length gradually decreases, with a significant reduction when the CAV penetration rate reaches 80%.
[0150] Table 4. Intersection signal timing schemes under different CAV penetration rates
[0151]
[0152] To verify the effectiveness of the proposed optimization model under different CAV penetration rates, this paper conducts a detailed analysis of intersection average vehicle delay. The main analysis results and findings are as follows:
[0153] (1) The impact of CAV penetration rate on average vehicle delay
[0154] This article analyzes the variation of average vehicle delay at intersections with CAV penetration rate, such as... Figure 3 As shown in the figure. The results show that: when the CAV penetration rate is 42%, the average vehicle delay calculated by setting up a dedicated CAV lane is the same as the classic model; when the CAV penetration rate is below 42%, the average vehicle delay calculated by setting up a dedicated CAV lane is greater than the classic model. Therefore, this intersection is not suitable for setting up dedicated CAV phases and dedicated approach lanes; as the CAV penetration rate increases, the advantages of the collaborative optimization model begin to emerge. When the CAV penetration rate is between 42% and 73%, the average vehicle delay calculated by setting up one dedicated CAV lane is less than that of the classic model and the case of setting up multiple dedicated CAV lanes, for example, when the CAV penetration rate is 50%, 60%, and 70%. When a dedicated CAV phase and one dedicated CAV approach lane are set up, the delay reduction rates are 7.3%, 19.1%, and 34.7% respectively compared to classic traffic organization and signal control schemes. This effectively reduces average vehicle delay and improves the service level of the intersection. When the CAV penetration rate is 73% or higher, the average vehicle delay calculated by setting up two dedicated CAV lanes is lower than that of the classic model and the case of setting up one dedicated CAV lane. Among them, when the CAV penetration rate is 80%, the average vehicle delay reduction rate is as high as 67%. This is because under high CAV penetration, the reasonable allocation of the right-of-way between CAV and HV is more effective in improving the intersection's capacity.
[0155] (2) Comparison of average vehicle delay for CAV and HV under different CAV penetration rates
[0156] This article compares the average vehicle delay of CAV and HV under different CAV penetration rates, such as Figure 4 As shown, when the CAV penetration rate is below 42%, no dedicated CAV lanes are deployed, and the average delay of CAV vehicles and the average delay of HV vehicles are equal to the average delay of the intersection. When the CAV penetration rate is above 42%, setting up dedicated CAV lanes and phases can reduce the average delay of the intersection. Furthermore, when the CAV penetration rate exceeds 70%, the reduction in the average delay of HV vehicles increases with the increase in CAV penetration rate.
[0157] In summary, the collaborative optimization model proposed in this paper demonstrates a more significant effect on optimizing traffic efficiency at mixed-traffic intersections under medium-to-high CAV penetration rates. Furthermore, as CAV penetration increases, the optimal lane organization scheme changes accordingly, and the average vehicle delay gradually decreases. The model effectively coordinates lane organization and signal control at intersections, thereby significantly improving the operational status of mixed-traffic intersections.
[0158] The present invention also provides a computer storage medium storing computer-readable instructions thereon, which, when executed by a computer, implement the above-described method.
[0159] Therefore, this invention employs the aforementioned method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, along with a computer storage medium. Through refined analysis of intersection traffic flow and combined with an intelligent optimization genetic algorithm, it coordinates and optimizes lane configuration, signal timing, and vehicle passage sequence at the intersection. By utilizing the autonomous driving capabilities of vehicles and vehicle-to-everything (V2X) communication technology, it achieves overall optimization of lane selection, passage timing, and speed guidance, effectively improving the intersection's traffic efficiency and operational stability. This provides strong theoretical support and practical guidance for the widespread application of autonomous driving technology in urban transportation.
[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.
Claims
1. A method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms, characterized in that, Includes the following steps: Step 1: Construct an objective function with average vehicle delay as the objective variable; Step 2: Based on traffic flow theory, set constraints and construct a collaborative optimization model for lane organization and signal control at mixed-traffic intersections; the constraints include lane layout constraints, traffic flow allocation constraints, and signal timing constraints. Step 3: Solve the objective function based on the genetic algorithm to obtain the optimal solution; initialize the model parameters, generate an initial population, design the signal timing scheme using the Webster method, define the fitness function, perform genetic operations on the samples, and iterate to obtain the scheme with the minimum average vehicle delay as the optimal solution. The specific constraints on lane layout are as follows: The layout of dedicated entrance lanes for autonomous driving follows the following formula: Where M is a sufficiently large positive number; sgn() is a sign function, which takes the value 1 when the parameter inside the function is positive, and 0 otherwise; The number of lanes for the dedicated autonomous driving lanes at the intersection i-entry direction; The green light time for the dedicated lane for autonomous driving at the intersection's i-entry direction; When the number of dedicated autonomous driving approach lanes is not zero, the dedicated autonomous driving phase at the intersection is activated. , When no dedicated approach lanes for automated driving are provided, the dedicated phase for automated driving at the intersection is closed. , ; For manual driving lanes, in the same approach direction, the constraints follow the following formula: ; in, These refer to the number of lanes at the intersection i-entry direction: left-turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane. The number of lane groups is guaranteed to be a non-negative integer, constrained by the following formula: ; In the formula, Indicates the number of lanes of type j in the direction of intersection i; For each direction of entry, the number of dedicated lanes for autonomous driving is equal to the number of dedicated lanes for autonomous driving deployed in that direction, and the number of regular lanes is equal to the number of entry lanes in that direction minus the number of dedicated lanes for autonomous driving. The number of lanes is defined by the following formula: ; ; ; in, These refer to the number of dedicated lanes for automated driving at the intersection's i-entry direction and the number of lanes for manual driving at the intersection's i-entry direction. Let i be the number of approach lanes in the direction of approach i at intersection i; i represents the direction of approach lane, and in a cross intersection i∈{1,2,3,4}, i represents the north approach lane, south approach lane, west approach lane, and east approach lane, respectively. Traffic flow allocation constraints include traffic flow constraints for dedicated lanes for autonomous driving, traffic volume constraints for lanes for manual driving, traffic flow balance constraints for turning lanes, traffic flow balance constraints for multi-turn lanes, and traffic flow distribution equilibrium constraints. The specific expressions are as follows: When a dedicated lane for autonomous driving exists, all autonomous vehicles use that lane; otherwise, the traffic flow is zero. The traffic flow constraint for the dedicated autonomous driving lane is as follows: as follows: ; In the formula, k is the flow direction symbol, k∈{L,T,R}, which represent left turn, straight ahead, and right turn, respectively; Traffic flow of autonomous vehicles in direction k of the i-import lane; Traffic volume constraints for manual driving lanes are as follows: ; in, Traffic volume using the manual driving lane in the direction of i entrance k; Traffic flow of manually driven vehicles in direction k of the i-entry lane; To ensure that all turning traffic flows at each approach lane are fully allocated to their respective lanes, the following formula is used to achieve traffic flow balance constraints at each turn: ; ; ; in, Let i be the traffic flow of lane j in the direction of k. Let be the traffic volume using the manual driving lane in direction k of entrance i, where k∈{L, T, R}. Let i represent the traffic volume of lane j with function at intersection i, where i∈{1,2,3,4} represents the north entrance lane, south entrance lane, west entrance lane, and east entrance lane, respectively; j represents the lane function, where j∈{1,2,...,6,7,8} represents the dedicated lane for automated driving, left turn lane, right turn lane, straight lane, left-straight combined lane, straight-right combined lane, left-right combined lane, and left-straight-right combined lane, respectively. To ensure that various types of entry lanes are correctly matched with the corresponding turning traffic flow, the following formula is used to achieve traffic flow balance constraints for multiple turning lanes: ; ; ; ; When the i-th approach lane has multiple lane groups facing a certain traffic flow, the traffic flow in that direction is distributed among the lane groups according to the principle of equal saturation. The traffic flow distribution balance constraint is achieved by the following formula: ; ; In the formula, It is an infinitesimal quantity; This refers to the time lost during vehicle startup. To determine the traffic capacity of the Class J functional lane in the direction of intersection i, This indicates the green light signal time for lane j in the direction of import i. This indicates the saturation level of lane j in the direction of import i. This indicates the number of lanes of type j1 in the direction of import i. This indicates the number of Class J2 lanes in the direction of import i. This indicates the saturation level of lane j1 in the direction of import i. This indicates the saturation level of lane j2 in the direction of import i. The signal timing constraints are as follows: Considering the weaving passage of autonomous vehicles, an additional dedicated autonomous driving phase is added to the existing signal phases. The design incorporates a "double-ring-three-group" structure for this dedicated autonomous driving phase. The autonomous driving traffic flows from each approach direction form the third dedicated autonomous driving phase. Specific constraints are as follows: The green light duration in signal timing is constrained by the following formula: ; in, The green light time for the Class J functional lane at the intersection i-th approach; and These are the minimum and maximum green light durations for the signal phase, respectively. The signal group needs to be determined in advance, and the signal flow to the green light time in the signal group should be equal, defined by the following formula: ; in, The green light duration for the m-th signal group; This is the m-th signal group; This indicates the green light signal time for lane j (i, j) in the direction of import i; The signal cycle duration T at the intersection is no greater than the maximum cycle duration. Not less than the minimum cycle duration The expression is as follows: ; ; In the formula, Time for clearing the intersection.
2. The method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithm according to claim 1, characterized in that, In step 1, the objective function is constructed as follows, using average vehicle delay as the objective variable: ; ; ; In the formula, d represents the average delay per vehicle, and Z is the set of vehicle types, Z = {A, H}, where A and H represent autonomous vehicles and manually driven vehicles, respectively. For the traffic volume of the Class J functional lane in the direction of intersection i approach, This indicates that the average vehicle delay is occurring in lane j of the i-direction approach. Let T represent the saturation level of lane j in the direction of intersection i, and T be the signal cycle duration. The green light ratio is the green light ratio of the Class J functional lane at the i-th approach of the intersection.
3. The method for coordinated optimization of lane organization and signal control at mixed-traffic intersections based on genetic algorithms according to claim 2, characterized in that, fitness function in step 3 as follows: ; In the formula, This indicates that all trains are delayed.
4. A computer storage medium, characterized in that: It stores computer-readable instructions that, when executed by a computer, implement the method as described in any one of claims 1 to 3.
Citation Information
Patent Citations
Intersection signal lamp timing optimization method based on improved Webster function and genetic algorithm
CN110517510A
Automatic / manual driving mixed intersection lane organization and signal optimization method
CN116863682A
