A traffic flow distribution method for a new type of mixed traffic flow

Abstract

The application discloses a kind of traffic flow distribution methods for new mixed traffic flow, the traffic data of city expressway distribution area is collected, data processing, model analysis and solution are carried out through steps S1 to S6, the distribution of HDVs and CAVs is realized;S1, traffic data processing;S2, distribution area scene construction;S3, upper model is constructed: the maximum distribution area operating efficiency;S4, lower model is constructed: mixed equilibrium distribution;S5, genetic algorithm solution algorithm;S6, output distribution area distribution result.The application proposes a kind of distribution area mixed traffic flow distribution method based on double-layer planning model, constructs the double-layer planning model for CAVs and HDVs mixed traffic flow equilibrium problem, analyzes the influence of distribution area main line lane layout and ramp flow on main line mixed traffic flow operation, reveals the dynamic characteristics and operating law of mixed traffic flow, solves and optimizes the optimal lane-changing path of CAVs, coordinates the equilibrium of mixed traffic flow, realizes the promotion of distribution area operation safety and efficiency.

Description

Technical Field

[0001] This invention relates to the fields of traffic engineering and artificial intelligence technology, and in particular to a traffic flow allocation method for diversion zones oriented towards novel mixed traffic flows. Background Technology

[0002] With the acceleration of urbanization and the continuous growth of transportation demand, urban transportation systems are facing increasingly complex challenges. Particularly in traffic diversion zones, mixed traffic flows of human-driven vehicles (HDVs) and connected and autonomous vehicles (CAVs) are becoming increasingly common, making traffic flow allocation even more complex. Traditional traffic flow allocation methods are often ineffective in handling such mixed traffic flows, leading to increased traffic congestion, low traffic efficiency, and increased safety hazards, severely restricting the overall operational efficiency of urban transportation systems.

[0003] In recent years, the rapid development of artificial intelligence technology has provided new solutions and technical means for the transportation field. Among them, bi-level programming models, as an efficient optimization tool, have demonstrated powerful capabilities in solving complex system optimization problems. Through a hierarchical optimization mechanism, it can simultaneously consider macro-efficiency objectives and micro-allocation behaviors, providing a new theoretical framework and technical path for solving mixed traffic flow allocation problems. Applying bi-level programming models to the mixed traffic flow allocation of CAVs and HDVs can not only effectively improve the traffic efficiency of diversion zones, but also provide a scientific basis for alleviating traffic congestion and optimizing resource allocation, possessing significant theoretical value and broad application prospects.

[0004] With the deep integration of intelligent transportation systems and artificial intelligence technologies, effectively optimizing the operational efficiency and safety of mixed traffic flow in diversion zones has always been an important research direction in the transportation field. As a key node in the transportation network, the layout of mainline lanes and ramp flow in diversion zones significantly impacts the operational status of mixed traffic flow. Especially in mixed traffic flow scenarios, traditional traffic management methods struggle to fully coordinate vehicle operating characteristics, leading to low traffic efficiency and increased safety hazards. Therefore, in-depth research into the operational patterns of mixed traffic flow in diversion zones and exploration of optimization strategies are of great significance for improving the overall performance of the transportation system. Summary of the Invention

[0005] To address the shortcomings of the existing technologies, this patent application aims to provide a traffic flow allocation method for diversion zones oriented towards novel mixed traffic flows. The method constructs a two-layer planning model to address the equilibrium problem of mixed traffic flows between CAVs and HDVs. By analyzing the impact of mainline lane layout and ramp flow on the operation of mixed traffic flows in the diversion zone, it reveals the dynamic characteristics and operational patterns of the mixed traffic flows, solves and optimizes the optimal lane-changing paths for CAVs, and coordinates the equilibrium of mixed traffic flows, thereby improving the safety and efficiency of diversion zone operations.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] A traffic flow allocation method for a new type of mixed traffic flow in a diversion zone is proposed. This method collects traffic data on the diversion zone of an urban expressway, processes the data and performs model analysis through steps S1 to S6, and then allocates HDVs and CAVs.

[0008] S1, Traffic Data Processing;

[0009] S2, Diversion Area Scene Construction;

[0010] S3. Construct the upper-level model: maximize the operating efficiency of the diversion zone;

[0011] S4. Construct the lower-level model: Hybrid balanced distribution;

[0012] S5. Genetic algorithm for solving the problem;

[0013] S6, Output the distribution results of the split zone.

[0014] Preferably, in step S1, traffic data processing includes data cleaning and data analysis;

[0015] The data cleaning process involves removing outliers, filling in missing values, and processing duplicate values ​​in the collected raw data.

[0016] Data analysis involves statistical analysis and mining of the cleaned data, using Tracker software to extract vehicle trajectories from drone videos, and employing interval cross-sections and lane division methods to analyze the spatial distribution characteristics and variation patterns of traffic volume, speed, headway, and lane-changing behavior.

[0017] Preferably, traffic data includes traffic flow, speed, and headway of HDVs and CAVs.

[0018] As a preferred option, in step S2, the area within 450 meters before the intersection point is defined as the diversion zone.

[0019] As a preferred option, in step S2, in the construction of the diversion zone scenario, the road network is constructed using SUMO software, a road network including main roads and diversion branch roads is designed, and the number of lanes, speed limits and traffic light parameters are set to simulate the real traffic environment.

[0020] Calibrate the following and lane-changing parameters for HDVs and CAVs respectively: For HDVs, use the Krauss following model and adjust the expected headway, maximum acceleration and driver reaction time parameters to simulate the uncertainty and conservatism of human driving behavior; For CAVs, use the CACC model and optimize following behavior based on the cooperative driving algorithm to shorten the headway and improve driving efficiency.

[0021] Regarding lane-changing behavior, both HDVs and CAVs adopted the LC2013 lane-changing model. By adjusting the lane-changing probability and driver aggression parameters in the LC2013 model, the impact of different driving behaviors on the stability and efficiency of traffic flow in the diversion zone was further studied.

[0022] As a preferred option, in step S3, under the premise of ensuring stable traffic flow, an upper-level model is constructed with the goal of maximizing operating efficiency, taking into account key indicators such as total traffic flow Q and total travel time T, as shown in formula (1).

[0023] maxZ=γQ-μT (1)

[0024] Introducing the headway parameter and combining it with the three factors of flow rate, density and speed, the dynamic characteristics of mixed traffic flow are shown in formula (2);

[0025]

[0026] The T-index improves the accuracy of describing traffic operation characteristics in the diversion zone under intelligent connected environment by integrating headway and lane change time, as shown in formula (3):

[0027]

[0028] In the formula: γ is the flow weight, μ is the time weight, and the two satisfy γ+μ=1; h is the average headway on lane i; n is the number of lanes, n∈[1,4]; t,i t1 is the sum of the headway distances between vehicles in lane i; e is the number of lane changes; l is the total number of lane changes; t1 is the time taken for CAVs and HDVs to change lanes once, with a value set between 5 and 10 seconds.

[0029] Based on the types of vehicles in front and behind, headway can be divided into four categories: CAVs following CAVs, CAVs following HDVs, HDVs following HDVs, and HDVs following CAVs, denoted by ρ1, ρ2, ρ3, and ρ4 respectively. Among them, the third and fourth categories, dominated by HDVs, have similar headway characteristics and can therefore be uniformly represented by ρ3. The relationship between the four categories of headway is 0 < ρ1 < ρ2 < ρ3 = ρ4, as shown in the following formula:

[0030]

[0031] In the formula: The number of CAVs with the first type of headway in lane i; The number of CAVs with the second type of headway in lane i; The number of HDVs for the headway of Class 3 vehicles in lane i; CAVs for lane i; For the HDV flow of lane i, satisfy

[0032] As a preferred option, in step S4, considering the differences in vehicle type and the impact of route selection on traffic allocation, a diversion correction factor m is proposed to improve the traditional BPR function, as shown in formulas (6)-(8):

[0033]

[0034] In the formula: t a t0 is the actual time required to traverse the road segment; q is the free travel time; C is the actual traffic volume on the road segment; a α represents the road segment capacity; β represents the undetermined parameters of the model, typically α = 0.15 and β = 4; L represents the road segment length; v0 represents the free-flow velocity; and m represents the diversion correction factor, which reflects the ratio of ramp flow to downstream flow. and The HDV and CAV flows that merge into ramp z on road segment a;

[0035] In mixed traffic flows, HDVs select the shortest travel time path, following the User Equilibrium (UE) principle; CAVs comprehensively consider the overall traffic situation, conforming to the System Optimal SO principle.

[0036] (1) UE Model

[0037] In the UE model, drivers choose the path that is most advantageous to themselves to maximize individual benefits, but this may lead to low overall efficiency. Based on this, a planning model for HDVs is constructed, with the specific formula as follows:

[0038]

[0039] In the formula, HDV flow rate on road segment a; Let r be the fixed CAV flow rate on road segment a; R be the starting point; S be the set of starting points; s be the destination; S be the set of destinations; k be the path; K be the destination. rs This is the set of feasible paths between OD pairs rs; Let OD be the HDV traffic of the k-th path of rs; For OD, the HDV requirement for the k-th path of rs; These are the variables related to the paths;

[0040] In this case, considering CAV flow as background flow, the BPR function of HDVs is adjusted from formula (6) to formula (14):

[0041]

[0042] (2) SO model

[0043] In an intelligent connected environment, CAVs (Conducting Access Control Systems) acquire real-time traffic information to optimize global traffic flow and travel time, thereby achieving optimal traffic allocation. Based on this, a planning model for CAVs is constructed, with the specific formula as follows:

[0044]

[0045] In the formula, The CAV flow rate on road segment a; The fixed HDV flow rate on road segment a; Let OD be the CAV traffic of the k-th path of rs; The CAV requirement of OD for the k-th path of rs;

[0046] In this case, considering HDV flow as background flow, the BPR function of CAVs should be adjusted from formula (6) to formula (20):

[0047]

[0048] Preferably, in step S5, during the solution process, the genetic algorithm encodes the solution to the problem into chromosomes by simulating the biological evolution process, and optimizes the population generation by generation using selection, crossover, and mutation operations. This step uses a genetic algorithm to solve the bilevel programming model, including the following steps:

[0049] P1: Parameter setting and population initialization: Set model parameters and key algorithm parameters, such as flow weight λ = 0.1, time weight μ = 0.9, population size 100, and maximum number of iterations 50; randomly generate the initial population, with each individual representing a path flow allocation scheme to ensure solution space diversity;

[0050] P2: Fitness Calculation and Selection: The overall fitness is calculated by fusing the results from the upper and lower layers. A higher value indicates a better allocation effect. Tournament selection is used to randomly select 5 individuals from the population and choose the one with the highest fitness as the parent to increase the retention probability of high-fit individuals.

[0051] P3: Crossover and Mutation Operations: Perform a two-point crossover operation to generate new individuals through gene information exchange, retaining superior characteristics and enhancing the diversity of the solution space; then randomly adjust the path flow distribution to increase population diversity, while setting the mutation probability to 0.2 to balance algorithm stability and search efficiency;

[0052] P4: Update and Termination Judgment: Update the population based on fitness values, retain superior individuals, and replace those with lower fitness with newly generated individuals; stop iteration when the maximum number of iterations is reached or the fitness changes tend to converge.

[0053] P5: Output the optimal solution: The final output is the optimized path traffic allocation scheme.

[0054] In summary, this patent proposes a traffic flow allocation method for diversion zones targeting novel mixed traffic flows. The rational allocation of mixed traffic flows is crucial for improving the efficiency and safety of diversion zones. Existing traffic assignment research often employs single-objective models, which struggle to adapt to complex traffic environments and dynamic optimization needs, and also have limitations in characterizing the complex traffic characteristics of diversion zones. To address these issues, this patent proposes a mixed traffic flow allocation method for diversion zones based on a two-level programming model. By constructing a multi-objective optimization model and improving the traditional BPR function, the model's adaptability and practical application value are significantly enhanced. The upper level maximizes operational efficiency by optimizing total flow and total travel time, while the lower level combines User Equilibrium (UE) and System Optimum (SO) principles to analyze the path allocation characteristics of HDVs and CAVs, comprehensively reflecting the traffic flow characteristics of mixed diversion zones. Furthermore, addressing the limitations of the traditional BPR function, a BPR function based on a diversion correction factor is proposed for the first time, fully considering the impact of diversion zone ramps on mainline flow, thus enhancing the model's adaptability to complex traffic environments. Attached Figure Description

[0055] Figure 1 This is a flowchart of a traffic flow allocation method for a novel mixed traffic flow, as described in this invention.

[0056] Figure 2 Map showing the impact range of the urban expressway diversion zone.

[0057] Figure 3 The flowchart shows how a genetic algorithm solves a bilevel programming model. Detailed Implementation

[0058] The present invention will now be described in further detail with reference to the accompanying drawings. In the description of the present invention, it should be understood that directional terms such as "upper," "lower," "top," and "bottom" indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. These terms are used only for the convenience of describing the present invention and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the scope of protection of the present invention. The directional terms "inner" and "outer" refer to the inner or outer contours relative to the outline of each component itself.

[0059] like Figure 1 As shown, a traffic flow allocation method for a new type of mixed traffic flow in a diversion zone is proposed. The method collects traffic data on the diversion zone of an urban expressway, and performs data processing and model analysis through steps S1 to S6 to achieve the allocation of HDVs and CAVs.

[0060] S1, Traffic Data Processing;

[0061] S2, Diversion Area Scene Construction;

[0062] S3. Construct the upper-level model: maximize the operating efficiency of the diversion zone;

[0063] S4. Construct the lower-level model: Hybrid balanced distribution;

[0064] S5. Genetic algorithm for solving the problem;

[0065] S6, Output the distribution results of the split zone.

[0066] In practice, step S1, traffic data processing includes data cleaning and data analysis.

[0067] The data cleaning process involves removing outliers, filling in missing values, and processing duplicate values ​​in the collected raw data to ensure the accuracy and integrity of the data.

[0068] Data analysis involves statistical analysis and mining of the cleaned data, using Tracker software to extract vehicle trajectories from drone videos, and employing interval cross-sections and lane division methods to analyze the spatial distribution characteristics and variation patterns of traffic volume, speed, headway, and lane-changing behavior.

[0069] In practice, traffic data includes traffic flow, speed, and headway of HDVs and CAVs.

[0070] In specific implementation, step S2 involves setting up the diversion zone scenario: In the Road Capacity Manual (HCM), this area is defined as the region within 450 meters upstream of the diversion point affected by diverted vehicles. Since there is currently a lack of a unified definition for urban expressway diversion zones, this patent refers to the standards in the HCM manual, such as... Figure 2 The affected area within 450 meters before the intersection point is designated as the diversion zone.

[0071] In specific implementation, in step S2, in the construction of the diversion zone scenario, the road network is constructed using SUMO (Simulation of Urban MObility) software. The road network, which includes main roads and diversion branch roads, is designed, and the number of lanes, speed limits, and traffic light parameters are set to simulate the real traffic environment.

[0072] Calibrate the following and lane-changing parameters for HDVs and CAVs respectively: For HDVs, use the Krauss following model and adjust the expected headway, maximum acceleration and driver reaction time parameters to simulate the uncertainty and conservatism of human driving behavior; For CAVs, use the CACC (Cooperative Adaptive Cruise Control) model and optimize following behavior based on cooperative driving algorithms to shorten headway and improve driving efficiency.

[0073] Regarding lane-changing behavior, both HDVs and CAVs adopted the LC2013 lane-changing model. This model comprehensively considers the lane-changing motivations from three aspects: safety, necessity, and rationality, and can simulate the lane-changing decision-making process of vehicles in the divergence zone. Through parameter calibration, it was ensured that HDVs and CAVs could realistically reflect their behavioral characteristics in the mixed traffic flow of the divergence zone, providing a reliable basis for subsequent simulation experiments. In addition, by adjusting the lane-changing probability and driver aggression parameters in the LC2013 model, the impact of different driving behaviors on the stability and efficiency of traffic flow in the divergence zone was further studied, providing data support for the optimization of traffic management and control strategies.

[0074] In specific implementation, step S3 involves constructing an upper-level model to maximize the operating efficiency of the diversion zone.

[0075] Under the premise of ensuring stable traffic flow, an upper-level model is constructed with the goal of maximizing operational efficiency, taking into account key indicators such as total traffic flow Q and total travel time T, as shown in formula (1);

[0076] maxZ=γQ-μT (1)

[0077] To address the problem that traditional traffic flow calculation methods based on the number of vehicles cannot fully reflect the complex operating characteristics of the diversion zone, the headway parameter is introduced. Combined with the three elements of flow, density and speed, the intrinsic relationship between the headway parameter and the flow distribution is analyzed in depth, so as to more accurately depict the dynamic characteristics of mixed traffic flow, as shown in formula (2).

[0078]

[0079] The T-index improves the accuracy of describing traffic operation characteristics in the diversion zone under intelligent connected environment by integrating headway and lane change time, as shown in formula (3):

[0080]

[0081] In the formula: γ is the flow weight, μ is the time weight, and the two satisfy γ+μ=1; h is the average headway on lane i; n is the number of lanes, n∈[1,4]; t,i t1 is the sum of the headway distances between vehicles in lane i; e is the number of lane changes; l is the total number of lane changes; t1 is the time taken for CAVs and HDVs to change lanes once, with a value set between 5 and 10 seconds.

[0082] Because CAVs are more intelligent, their headway is usually shorter, while HDVs are affected by driver behavior and have a larger headway. Based on the type of vehicle in front and behind, headway can be divided into four categories: CAVs following CAVs, CAVs following HDVs, HDVs following HDVs, and HDVs following CAVs, denoted by ρ1, ρ2, ρ3, and ρ4 respectively. The third and fourth categories, dominated by HDVs, have similar characteristics and can therefore be uniformly represented by ρ3. The relationship between the four types of headway is 0 < ρ1 < ρ2 < ρ3 = ρ4, as shown in the following formula:

[0083]

[0084] In the formula: The number of CAVs with the first type of headway in lane i; The number of CAVs with the second type of headway in lane i; The number of HDVs for the headway of Class 3 vehicles in lane i; CAVs for lane i; For the HDV flow of lane i, satisfy

[0085] In specific implementation, in step S4, the lower-level model: hybrid equilibrium allocation:

[0086] Considering the differences in vehicle type and the impact of route selection on traffic flow allocation, a diversion correction factor m is proposed to improve the traditional BPR function, as shown in formulas (6)-(8), to more accurately reflect the traffic flow characteristics of the diversion zone:

[0087]

[0088] In the formula: t a t0 is the actual time required to traverse the road segment; q is the free travel time; C is the actual traffic volume on the road segment; aα represents the road segment capacity; β represents the undetermined parameters of the model, typically α = 0.15 and β = 4; L represents the road segment length; v0 represents the free-flow velocity; and m represents the diversion correction factor, which reflects the ratio of ramp flow to downstream flow. and The HDV and CAV flows that merge into ramp z on road segment a;

[0089] In mixed traffic flows, HDVs and CAVs exhibit significant differences in path selection, and a uniform allocation principle is insufficient to effectively reflect their characteristics. HDVs typically select the shortest travel time path, adhering to the User Equilibrium (UE) principle; CAVs comprehensively consider the overall traffic situation, conforming to the System Optimal Sorting (SO) principle.

[0090] (1) UE Model

[0091] In the UE model, drivers choose the path that is most advantageous to themselves to maximize individual benefits, but this may lead to low overall efficiency. Based on this, a planning model for HDVs is constructed, with the specific formula as follows:

[0092]

[0093] In the formula, HDV flow rate on road segment a; Let r be the fixed CAV flow rate on road segment a; R be the starting point; S be the set of starting points; s be the destination; S be the set of destinations; k be the path; K be the destination. rs This is the set of feasible paths between OD pairs rs; Let OD be the HDV traffic of the k-th path of rs; For OD, the HDV requirement for the k-th path of rs; These are the variables related to the paths;

[0094] In this case, considering CAV flow as background flow, the BPR function of HDVs is adjusted from formula (6) to formula (14):

[0095]

[0096] (2) SO model

[0097] In an intelligent connected environment, CAVs (Conducting Access Control Systems) acquire real-time traffic information to optimize global traffic flow and travel time, thereby achieving optimal traffic allocation. Based on this, a planning model for CAVs is constructed, with the specific formula as follows:

[0098]

[0099] In the formula, The CAV flow rate on road segment a; The fixed HDV flow rate on road segment a; Let OD be the CAV traffic of the k-th path of rs; The CAV requirement of OD for the k-th path of rs;

[0100] In this case, considering HDV flow as background flow, the BPR function of CAVs should be adjusted from formula (6) to formula (20):

[0101]

[0102] In specific implementation, step S5 involves using a genetic algorithm to solve the problem.

[0103] In the solution process, the genetic algorithm simulates the biological evolution process, encoding the solution to the problem as chromosomes, and optimizes the population generation by generation using selection, crossover, and mutation operations. This step uses a genetic algorithm (GA) to solve the bilevel programming model, including the following steps, the specific process of which is as follows: Figure 3 As shown:

[0104] P1: Parameter setting and population initialization: Set model parameters and key algorithm parameters, such as flow weight λ = 0.1, time weight μ = 0.9, population size 100, and maximum number of iterations 50; randomly generate the initial population, with each individual representing a path flow allocation scheme to ensure solution space diversity;

[0105] P2: Fitness Calculation and Selection: The overall fitness is calculated by fusing the results from the upper and lower layers. A higher value indicates a better allocation effect. Tournament selection is used to randomly select 5 individuals from the population and choose the one with the highest fitness as the parent to increase the retention probability of high-fit individuals.

[0106] P3: Crossover and Mutation Operations: Perform a two-point crossover operation to generate new individuals through gene information exchange, retaining superior characteristics and enhancing the diversity of the solution space; then randomly adjust the path flow distribution to increase population diversity, while setting the mutation probability to 0.2 to balance algorithm stability and search efficiency;

[0107] P4: Update and Termination Judgment: Update the population based on fitness values, retain superior individuals, and replace those with lower fitness with newly generated individuals; stop iteration when the maximum number of iterations is reached or the fitness changes tend to converge.

[0108] P5: Output the optimal solution: The final output is the optimized path traffic allocation scheme.

[0109] In step S6, the distribution results of the split-flow area are output:

[0110] Through steps S1-S5, a bi-level planning model for the diversion zone based on mixed traffic flow allocation is obtained, and a genetic algorithm is used to solve it, resulting in the mixed traffic flow allocation of the diversion zone. This model can be used to study various types of traffic data in the diversion zone.

[0111] This invention proposes a bi-level programming model for the allocation of mixed traffic flow in diversion zones. The aim is to construct a bi-level programming model for the equilibrium problem of mixed traffic flow between CAVs and HDVs. By analyzing the impact of the mainline lane layout and ramp flow on the operation of mixed traffic flow in the diversion zone, the dynamic characteristics and operation rules of mixed traffic flow are revealed. The optimal lane-changing path of CAVs is solved and optimized to coordinate the equilibrium of mixed traffic flow, thereby improving the safety and efficiency of diversion zone operation.

[0112] Finally, it should be noted that those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A traffic flow allocation method for a new type of mixed traffic flow, characterized in that, Traffic data on the urban expressway diversion zone is collected, and data processing and model analysis are performed through steps S1 to S6 to achieve the allocation of HDVs and CAVs. S1, Traffic Data Processing; S2, Diversion Area Scene Construction; S3. Construct the upper-level model: maximize the operating efficiency of the diversion zone; S4. Construct the lower-level model: Hybrid balanced distribution; S5. Genetic algorithm for solving the problem; S6. Output the distribution results of the splitter zone; In step S3, under the premise of ensuring stable traffic flow, an upper-level model is constructed with the goal of maximizing operating efficiency, taking into account key indicators such as total traffic flow Q and total travel time, as shown in formula (1); （1） Introducing the headway parameter and combining the three elements of flow rate, density and speed, the dynamic characteristics of mixed traffic flow are shown in formula (2); （2） The T-index improves the accuracy of describing traffic operation characteristics in the diversion zone under intelligent connected environment by integrating headway and lane change time, as shown in formula (3): （3） In the formula: For traffic weight, As time weights, both satisfy ; For lane Average headway; For the number of lanes, ; For lane The sum of the distances when the vehicle is mounted; This refers to the number of lane changes. This represents the total number of lane changes. The time taken for CAVs and HDVs to switch lanes once is set between 5 and 10 seconds. Based on the types of vehicles in front and behind, the headway can be divided into four categories: CAVs following CAVs, CAVs following HDVs, HDVs following HDVs, and HDVs following CAVs, respectively denoted by... , , , This indicates that the headway of Category III and Category IV vehicles, dominated by HDVs, has similar characteristics, therefore they can be uniformly represented using... This indicates the relationship between the headway distances of the four types of vehicles. The specific formula is as follows: （4） （5） In the formula: For lane The number of CAVs at the headway of the first type of vehicle; For lane The number of CAVs at the headway of the second type of vehicle; For lane The number of HDVs at the headway of the third type of vehicle; For lane CAVs; For lane HDV traffic, to meet ; In step S4, considering the differences in vehicle type and the impact of route selection on traffic allocation, a diversion correction factor is proposed to improve the traditional BPR function, as shown in formulas (6)-(8): （6） （7） （8） In the formula: This refers to the actual time required to traverse the road segment; Free time for driving; To improve the traffic capacity of the road section; , These are the undetermined parameters for the model, typically α=0.15 and β=4; The length of the road segment; For free flow velocity; This is a diversion correction factor, used to reflect the ratio of ramp flow to downstream flow; and The HDV and CAV flows that merge into ramp z on road segment a; In mixed traffic flows, HDVs select the shortest travel time path, following the User Equilibrium (UE) principle; CAVs comprehensively consider the overall traffic situation, conforming to the System Optimal SO principle. (1) UE model In the UE model, drivers choose the path that is most advantageous to themselves to maximize individual benefits, but this may lead to low overall efficiency. Based on this, a planning model for HDVs is constructed, with the specific formula as follows: （9） （10） （11） （12） （13） In the formula, For road section Upload HDV traffic; For road section Install a fixed CAV flow rate; Starting point; The starting set; End point; For the set of endpoints; For path; For OD The set of feasible paths between them; For OD No. HDV traffic for each path; For OD No. HDV requirements for each path; These are the variables related to the paths; In this case, considering CAV flow as background flow, the BPR function of HDVs is adjusted from formula (6) to formula (14): （14） (2) SO model In an intelligent connected environment, CAVs (Conducting Access Control Systems) acquire real-time traffic information to optimize global traffic flow and travel time, thereby achieving optimal traffic allocation. Based on this, a planning model for CAVs is constructed, with the specific formula as follows: （15） （16） （17） （18） In the formula, For road section Increase CAV traffic; For road section HDV fixed flow rate; For OD No. CAV traffic for each path; For OD No. CAV requirements for each path; In this case, considering HDV flow as background flow, the BPR function of CAVs should be adjusted from formula (6) to formula (20): （19） （20）。 2. The traffic flow allocation method for a new type of mixed traffic flow according to claim 1, characterized in that, In step S1, traffic data processing includes data cleaning and data analysis; The data cleaning process involves removing outliers, filling in missing values, and processing duplicate values ​​in the collected raw data. Data analysis involves statistical analysis and mining of the cleaned data, using Tracker software to extract vehicle trajectories from drone videos, and employing interval cross-sections and lane division methods to analyze the spatial distribution characteristics and variation patterns of traffic volume, speed, headway, and lane-changing behavior.

3. The traffic flow allocation method for a new type of mixed traffic flow according to claim 2, characterized in that, Traffic data includes traffic flow, speed, and headway for both HDVs and CAVs.

4. The traffic flow allocation method for a new type of mixed traffic flow according to claim 1, characterized in that, In step S2, the affected area within 450 meters before the intersection point is defined as the diversion zone.

5. A traffic flow allocation method for a new type of mixed traffic flow according to claim 1, characterized in that, In step S2, during the construction of the diversion zone scenario, the road network is constructed using SUMO software. A road network including main roads and diversion branch roads is designed, and the number of lanes, speed limits, and traffic light parameters are set to simulate the real traffic environment. Calibrate the following and lane-changing parameters for HDVs and CAVs respectively: For HDVs, use the Krauss following model and adjust the expected headway, maximum acceleration and driver reaction time parameters to simulate the uncertainty and conservatism of human driving behavior; For CAVs, use the CACC model and optimize following behavior based on the cooperative driving algorithm to shorten the headway and improve driving efficiency. Regarding lane-changing behavior, both HDVs and CAVs adopted the LC2013 lane-changing model. By adjusting the lane-changing probability and driver aggression parameters in the LC2013 model, the impact of different driving behaviors on the stability and efficiency of traffic flow in the diversion zone was further studied.

6. The traffic flow allocation method for a new type of mixed traffic flow according to claim 1, characterized in that, In step S5, during the solution process, the genetic algorithm encodes the solution to the problem into chromosomes by simulating the biological evolution process, and optimizes the population generation by generation using selection, crossover, and mutation operations. This step uses a genetic algorithm to solve the bilevel programming model, including the following steps: P1: Parameter setting and population initialization: Set model parameters and key algorithm parameters, such as flow weight λ=0.1, time weight μ=0.9, population size 100, and maximum number of iterations 50; randomly generate the initial population, with each individual representing a path flow allocation scheme to ensure solution space diversity; P2: Fitness Calculation and Selection: The overall fitness is calculated by fusing the results from the upper and lower layers. A higher value indicates a better allocation effect. Tournament selection is used to randomly select 5 individuals from the population. The individual with the highest fitness is selected as the parent to increase the retention probability of high-fit individuals. P3: Crossover and mutation operations: Perform a two-point crossover operation to generate new individuals through gene information exchange, retaining superior characteristics and enhancing the diversity of the solution space; then randomly adjust the path flow distribution to increase population diversity, while setting the mutation probability to 0.2 to balance the stability of the algorithm and search efficiency; P4: Update and Termination Judgment: Update the population based on fitness values, retain superior individuals, and replace those with lower fitness with newly generated individuals; stop iteration when the maximum number of iterations is reached or the fitness changes tend to converge. P5: Output the optimal solution: The final output is the optimized path traffic allocation scheme.

Images