A two-layer collaborative control method considering traffic risks in expressway merging areas under connected environments

Abstract

The present invention discloses a dual-layer collaborative control method that considers traffic risks in expressway merging areas in a networked environment. The method is applicable to expressway merging areas and includes the following steps: 1. predicting the traffic flow state for the next time period and calculating the accident risk and congestion risk in the merging area; 2. activating the dual-layer collaborative control strategy when the traffic risk exceeds the corresponding threshold and the ramp queue length exceeds the threshold, obtaining the optimal control parameters for the next time period based on the upper and lower layer objectives and optimization algorithms, and implementing collaborative control; 3. activating the ramp regulation control strategy when the risk exceeds the threshold and the ramp queue length does not exceed the threshold, calculating the optimal ramp regulation rate based on the upper layer objective. The present invention takes into account both reducing accident risks and alleviating congestion, establishing a dual-layer planning model, and enabling the upper and lower layer control strategies to coordinate and match, balancing the network traffic distribution, and overall realizing intelligent collaborative management of traffic flow, which can significantly improve the safety, stability, and traffic efficiency of complex road sections.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent connected driving applications, specifically a two-layer collaborative control method that considers traffic risks in expressway merging areas under a connected environment; Background Art

[0002] With the development of intelligent and information-based technologies, the application of the Internet of Vehicles (IoV) is becoming increasingly widespread. In complex urban road environments, particularly in areas where expressway mainlines merge with ramps, the risk of vehicle collisions is high, and these areas are also prone to traffic congestion. To improve road capacity and safety, intelligent traffic management and control technologies are needed.

[0003] Traditional traffic control methods for expressway merging areas often rely on static rules, making them inflexible in responding to complex and changing traffic conditions. In recent years, various adaptive and collaborative traffic management methods have emerged, leveraging connected vehicle (IoV) technology. These methods collect traffic flow parameters through wireless communications between vehicles and roadside monitoring equipment, analyze and assess traffic conditions, and then adjust traffic signal timings or speed limits to improve traffic flow and reduce accident risks. However, existing methods often rely on single-level control, inadequately considering the combined traffic risks of the mainline and ramps, resulting in a relatively crude and inadequate assessment and control of traffic conditions. To achieve refined control across different road sections, it is necessary to establish a multi-level traffic flow model and employ collaborative control methods at different levels to ensure targeted and real-time control strategies. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the present invention adopts a two-layer control strategy based on the main line and ramps, and combines a numerical optimization algorithm as an optimization tool to provide a two-layer collaborative control method that takes into account the traffic risks of expressway merging areas in a networked environment, in order to achieve comprehensive management of expressway merging areas with self-adaptation and global optimization. The two-layer planning model establishes two-level planning models, each with multiple objectives, and performs step-by-step optimization. It can decompose the problem more clearly, model the two sub-problems of the main line and the ramp separately, and reduce the complexity of the problem. The two-layer collaborative control method enables the merging area to balance the collision risk and congestion risk in the collaborative optimization of the upper main line speed limit and the lower ramp vehicle control rate, quickly search for the global optimal parameter solution, and provide a broader optimization space for the control strategy. It can better balance the traffic flow distribution, reduce the collision risk and alleviate congestion, improve the safety and efficiency of the traffic system, and provide an efficient and feasible solution for the intelligent optimization of merging area traffic.

[0005] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0006] The present invention provides a double-layer cooperative control method for considering traffic risks in expressway merging areas in a networked environment. The method is characterized in that the expressway merging area refers to the intersection area of ​​the main line area and the ramp area of ​​the expressway; taking the vehicle driving direction as the positive direction and the length of the merging area as the standard length, the upstream area of ​​the merging area is divided into a main line cells according to the standard length, the merging area is used as a merging cell and recorded as the a+1th main line cell, and the downstream area of ​​the merging area is divided into b main line cells according to the standard length, thereby obtaining a+b+ 1 mainline cell; according to the vehicle travel direction, any mainline cell among the a+b+1 mainline cells is recorded as the i-th mainline cell, the ramp area is divided into J ramp cells according to the standard length, and the ramp cell adjacent to the a+1-th mainline cell is taken as the J-th ramp cell, and any ramp cell is defined as the j-th ramp cell; the n mainline cells upstream of the a-th mainline cell are set as variable speed limit control implementation cells, and the J-th ramp cell is set as the ramp control implementation cell, 1≤n≤a-1; the method is performed in the following steps:

[0007] Step 1: Use roadside intelligent devices to obtain traffic flow parameters of each cell on the expressway mainline and ramp in time period t, and use them to predict traffic flow parameters in time period t+1;

[0008] Step 2: Input the traffic flow parameters of the t+1 time period into the accident risk assessment model and the congestion risk assessment model respectively, and calculate the accident risk value CR of the merging cell in the t+1 time period. a+1,m (t+1) and congestion risk value JAM a+1,m (t+1); inputting the traffic flow parameters of the t+1 time period into the ramp queue length model, and calculating the queue length W(t+1) of ramp cell J in the t+1 time period;

[0009] Step 3: Determine CR a+1,m (t+1) Whether the accident risk threshold Δ is exceeded CR or JAM a+1,m (t+1) Whether the congestion risk threshold Δ is exceeded JAM If yes, go to step 4, otherwise go to step 10;

[0010] Step 4: Based on the various control objectives of the mainline and ramps, a two-level planning model for the merging area in time period t+1 is constructed. This model includes an upper-level planning model with the control objectives of minimizing the accident risk and maximizing the traffic turnover in the merging area, and a lower-level planning model with the control objectives of minimizing the average waiting time and energy consumption of ramp vehicles.

[0011] Step 5: Determine whether the ramp queue length W(t+1) exceeds the threshold Δ W, if yes, then execute the coordinated control strategy of step 6, otherwise, execute the ramp control strategy of step 7;

[0012] Step 6: Iteratively solve the merging area bi-level programming model using a hybrid genetic simulated annealing algorithm to determine the optimal variable speed limit and the optimal ramp adjustment rate in the t+1 time period, and then execute step 8;

[0013] Step 7: Iteratively solve the upper-level planning model in the merging area bi-level planning model using a genetic algorithm to determine the optimal ramp regulation rate in the t+1 time period, and then execute step 9;

[0014] Step 8: In the t+1 time period, the variable speed limit control implementation cell implements coordinated control according to the optimal variable speed limit value, and the ramp control cell implements coordinated control according to the optimal ramp adjustment rate, and then executes step 10;

[0015] Step 9: In the t+1 time period, the ramp control cell performs ramp control according to the optimal ramp regulation rate, and executes step 10;

[0016] Step 10: After assigning t+1 to t, if t reaches the time threshold T, stop the control; otherwise, return to step 1 and execute sequentially.

[0017] The dual-layer collaborative control method of the present invention is also characterized in that step 1 includes:

[0018] Step 1.1: Calculate the sending and receiving capabilities of each cell in time period t:

[0019] According to formula (1) to formula (2), the sending capacity σ of the i-th main line cell in the t-th time period is calculated i,m (t) and receiving capacity δ i,m (t):

[0020] σ i,m (t) = min{v i,m (t)·k i,m (t)·n i,m ,Q m ·n i,m}·Δt (1)

[0021] δ i,m (t) = {w m ·(k jam,m -k i,m (t))·n i,m ,Q m ·n i,m}·Δt (2)

[0022] In formula (1) to formula (2), v i,m (t), ki,m (t) represents the average velocity and average density of the i-th main line cell in the t-th time period; k jam,m , w m , Q m , n i,m They represent the mainline blocking density, wave speed, single lane maximum capacity and number of lanes of the i-th mainline cell respectively, and Δt is the interval between adjacent time periods;

[0023] According to formula (3) to formula (4), the sending capacity σ of the j-th ramp cell in the t-th time period is calculated j,r (t) and receiving capacity δ j,r (t);

[0024] σ j,r (t) = min{v j,r (t)·k j,r (t)·n j,r ,Q r ·n j,r}·Δt (3)

[0025] δ j,r (t) = {w r ·(k jam,r -k j,r (t))·n j,r ,Q r ·n j,r}·Δt (4)

[0026] In formula (3) to formula (4), v j,r (t), k j,r (t) represents the average speed and average density of the j-th ramp cell in the t-th time period, k jam,r , w r , Q r , n j,r They represent the ramp blocking density, wave speed, maximum single lane capacity and number of lanes of the j-th ramp cell respectively;

[0027] Step 1.2: Calculate the transmission flow q of the i-th mainline cell in the t-th time period according to formula (5): i,m (t);

[0028]

[0029] In formula (5), q i,m (t) represents the traffic volume transmitted downstream by the i-th mainline cell in the t-th time period, p on (t) represents the confluence ratio between the mainline area and the ramp area in the tth time period; σ a,m (t) represents the sending capacity of the ath mainline cell in the tth time period, δa+1,m (t) represents the receiving capacity of the a+1th mainline cell in the tth time period, σ J,r (t) represents the sending capacity of the J-th ramp cell in the t-th time period;

[0030] Step 1.3: Predict the transmission flow q of the jth ramp cell in the tth time period according to formula (6): j,r (t);

[0031]

[0032] In formula (6), σ J,r (t) represents the sending capacity of the J-th ramp cell in the t-th time period; δ j+1,r (t) represents the receiving capacity of the j+1th ramp cell in the tth time period;

[0033] Step 1.4: Predict the average density k of the i-th main line cell in the t+1 time period according to equations (7) and (8). i,m (t+1) and the average density k of the j-th ramp cell j,r (t+1):

[0034]

[0035] k j,r (t+1)=k j,r (t)+(q j-1,r (t)-q j,r (t))×Δt / l j,r (8)

[0036] In formula (7) to formula (8), l i,m represents the length of the i-th main line cell, l j,r represents the length of the j-th ramp cell;

[0037] Step 1.5: Predict the average velocity v of the i-th main line cell in the t+1 time period according to equations (9) and (10): i,m (t+1) and the average velocity v of the j-th ramp cell j,r (t+1);

[0038]

[0039]

[0040] In formula (9) to formula (10), V f,m represents the free flow velocity of the main line cell, w m represents the congestion wave propagation speed of the main line cell, k cm represents the critical density of the main line cell, V f,rrepresents the free flow velocity of the ramp cell, w r represents the congestion wave propagation speed of the ramp cell, k cr represents the critical density of ramp cells.

[0041] In step 2, the accident risk assessment model is constructed using formula (11), the congestion risk assessment model is constructed using formula (12), and the ramp queue length model is constructed using formula (13):

[0042]

[0043]

[0044] W(t+1)=W(t)+(q J,r (t)-q J-1,r (t))·Δt (13)

[0045] In formula (11)-(13), x1, x2, x3, x4, x5 are the coefficients of five variables, β is the intercept, and v a-3,m (t+1) represents the average velocity of the a-3th main line cell in the t+1th time period, k a-2,m (t+1) represents the average density of the a-2th main line cell in the t+1th time period, v a+1,m (t+1) represents the average velocity of the a+1th main line cell in the t+1th time period, q a+2,m (t) represents the transmission flow of the a+2th mainline cell in the tth time period, dq a-3,a+2,m V represents the difference in transmission flow between the a-3th mainline cell and the a+2th mainline cell in the tth time period, max Indicates the maximum speed limit on the expressway, v a+1,m (t+1) represents the average velocity of the a+1th main line cell in the t+1th time period, q a+1,m (t) represents the transmission flow of the a+1th mainline cell in the tth time period, Q m represents the maximum traffic capacity of a single lane, q J-1,r (t) represents the transmission flow of the J-1th ramp cell upstream of ramp cell J in the tth time period, W(t) represents the number of vehicles queuing in ramp cell J in the tth time period. When t = 1, the number of vehicles queuing in the Jth ramp cell in the tth time period, W(t), is detected by the roadside intelligent device. Δt is the interval between adjacent time periods.

[0046] In step 4, the upper-level planning model is constructed using formula (14) and the lower-level planning model is constructed using formula (15):

[0047]

[0048] In formula (14)-formula (15), J U (t+1) represents the objective function value of the upper-level planning model in the t+1 time period, J D (t+1) represents the objective function value of the lower-level planning model in the t+1th time period, μ1, μ2, μ3, α1 and α2 are five-dimensional balance coefficients, l a+1,m represents the length of the a+1th main line cell, C D is the drag coefficient, A is the projected area in the direction of the car's travel, ρ is the air density, Z is the car's gravity, f is the rolling resistance coefficient, and v J,r (t+2) represents the average speed of the ramp control cell in the t+2 time period;

[0049] Formula (16) is used to construct the constraints of the two-level planning model of the merging area:

[0050]

[0051] In formula (16), v i (t) represents the variable speed limit value of the i-th variable speed limit control cell in the t-th time period, v i-1 (t) represents the variable speed limit value of the i-1th variable speed limit control cell in the tth time period, v i (t+1) represents the variable speed limit value of the i-th variable speed limit control cell in the t+1 time period, spd diff,s Indicates the speed limit difference threshold of adjacent variable speed limit control cells in the same time period, spd diff,t represents the speed limit difference threshold of the same variable speed limit control cell in adjacent time periods, v min Indicates the minimum speed limit on the expressway, v max Indicates the maximum speed limit on the expressway.

[0052] The hybrid genetic simulated annealing algorithm in step 6 is solved according to the following steps:

[0053] Step 6.1, define the population size as M, the current number of evolutions as G, and the maximum number of evolution generations as G max , the population crossover probability is p c , the population mutation probability is p v , initialize G = 1;

[0054] Step 6.2: Generate the G-generation genetic population in the t+1 time period

[0055] Step 6.2.1: Generate M chromosomes in the t+1 time period according to formula (16) and form the G-th generation variable speed-limited population. in, Represents the G-th generation variable speed limit population in the t+1 time period The sth chromosome in , and each chromosome represents a variable speed limit control strategy, Represents the sth variable speed limit control strategy The variable speed limit value of the e-th variable speed limit control cell in ; e=1,2,...,n,s=1,2,...,M;

[0056] Step 6.2.2: Randomly generate the G-th generation variable speed limit population The sth chromosome P s G (t+1) ramp regulation rate real number sub-gene string in, express The corresponding s-th ramp regulation rate sub-gene in , and Thus, the G generation genetic population is obtained in, represents the sth individual, s=1,2,...,M;

[0057] Step 6.3: Obtain the G-generation genetic population The sth individual The corresponding minimum fitness value Ramp regulatory subgene Thus, the sth chromosome and Form the sth new individual, thus reaching the Gth generation genetic population The Gth generation new genetic population is composed of the chromosome of each individual and the ramp adjustment rate sub-gene of its minimum fitness value

[0058] Step 6.3.1, the G generation genetic population The sth individual Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the traffic flow state quantity predicted by each cell of the Gth generation in the t+2th time period;

[0059] Step 6.3.2: Substitute the traffic flow state predicted by each cell in the t+2th time period into the lower-level planning model, and use the inverse of the lower-level planning model as the fitness function to obtain the Gth generation genetic population. The sth individual The fitness set in, express The corresponding s-th ramp regulation rate sub-gene The fitness value of

[0060] Step 6.3.3: Genetic population of generation G The sth individual The fitness value set Compare the fitness values ​​of each ramp regulation rate sub-gene in the and select the minimum fitness value The corresponding ramp regulation rate sub-gene is recorded as

[0061] Step 6.4, G generation genetic population The chromosomes of each individual and the ramp regulation rate sub-genes of each individual are crossover and mutation operations are performed to generate the offspring population of the G generation population. For offspring population Perform selection operation to obtain the genetic population of the G+1 generation

[0062] Step 6.5: Calculate the genetic population of the G+1 generation according to the process of step 6.3 The minimum fitness value corresponding to all individuals in in, express The minimum fitness value corresponding to the Mth individual in The sth chromosome P in s G+1 (t+1) and The corresponding ramp regulation rate sub-gene Forming the G+1 generation new genetic population The sth individual;

[0063] like Less than Then retain the G+1 generation new genetic population The sth individual Otherwise, Assign to Will Assign to the sth new individual Thus updating the G+1 generation new genetic population The sth new individual;

[0064] Step 6.6, assign G+1 to G, and judge whether G<G max Is it true? If so, go to step 6.4; otherwise, it means that G is completed. max The Gth iteration is obtained max New genetic population

[0065] Step 6.7, Performing an annealing operation;

[0066] Step 6.7.1, define the initial temperature as H0 and the final temperature as Hend , temperature drop coefficient α and maximum number of cycles Inter; define the current temperature as H k , and initialize H k =H0; the current number of cycles is inter, and initialize inter=1; As the current inter generation population, denoted as

[0067] Step 6.7.2, Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the traffic flow state quantity predicted by each inter-generation cell in the t+2 time period. The inverse of formula (14) is used as the fitness function to obtain The upper fitness value of each individual in the inter generation is selected, and the individual corresponding to the minimum fitness value is recorded as the optimal individual of the inter generation. in, express Minimum fitness value The corresponding chromosomes express The corresponding ramp regulation rate sub-gene;

[0068] Step 6.7.3: According to the Metropolis criterion and H k ,right Each individual is selected to obtain the inter+1 generation population

[0069] Step 6.7.4: Determine whether inter = Inter. If so, proceed to step 6.7.5. Otherwise, assign inter + 1 to inter and proceed to step 6.7.2.

[0070] Step 6.7.5, if H k =H end , which means the optimal variable speed limit value and the optimal ramp adjustment rate in the t+1 time period are obtained Otherwise, αH k Assign to H k , proceed to step 6.7.3.

[0071] Step 7 is to obtain the optimal ramp adjustment rate in the t+1 time period when executing the ramp control strategy according to the following steps:

[0072] Step 7.1. Initialize the population size M, the current number of evolutions G, and the maximum number of evolution generations G max , population crossover probability p c , population mutation probability p v , initialize G = 1;

[0073] Step 7.2: Randomly generate M chromosomes in the t+1 time period. Each chromosome represents a ramp adjustment rate. The G-th iteration set is obtained as And r s (t+1)∈[0,1], s=1,2,...,M;

[0074] Step 7.3: Based on the improved cellular transmission model, predict the traffic flow state of each cell in the t+2 time period under the ramp adjustment rate corresponding to the sth chromosome in the t+1 time period;

[0075] Step 7.4: Substitute the traffic flow state quantity of the t+1th time period corresponding to the sth chromosome in the Gth generation population into the upper planning model, and use the inverse of the upper planning model formula (14) as the fitness function to obtain the fitness set of the Gth generation population. It represents the corresponding fitness value of the sth chromosome of the Gth generation under the objective function in the t+1th time period;

[0076] Step 7.5: Set the fitness values ​​of all chromosomes for the Gth time Compare the fitness values ​​of each chromosome in the G generation and save the chromosome with the minimum fitness value of each group This value indicates that this strategy is to minimize the traffic risk value of the merging area in the t+2 time period. The minimum fitness value is recorded as

[0077] Step 7.6: Set the Gth generation population to a population size of M Perform crossover and mutation to generate the G generation population The offspring population is Its population size is also M, and the offspring population Perform selection operations to keep the population size M, and the resulting chromosomes serve as the parent population of the G+1 generation. s=1,2,...,M;

[0078] Step 7.7: Calculate the parent population of the G+1 generation according to steps 7.3 to 7.4. The fitness value set of all chromosomes in The minimum fitness value of each group of chromosomes in the G+1 generation is The corresponding chromosome is denoted as like Less than in step 7.5 Then the updated minimum fitness value is And save its corresponding chromosome; otherwise, use the chromosome corresponding to the minimum fitness value in the G generation chromosome Replace the chromosome with the minimum fitness value in the G+1 generation chromosome and enter the next generation;

[0079] Step 7.8: Assign G+1 to G and determine whether G < G max Is it true? If so, go to step 7.6; otherwise, it means that G is completed. max The final population set is recorded as The population in The gene is the optimal solution, which is the optimal ramp adjustment ratio ratio(t+1);

[0080] The improved cellular transmission model predicts the traffic flow state of any generation in the t+2 time period according to the following steps:

[0081] Step a: record the chromosome of any individual in any generation of population as (v a-n,m,vsl (t+1),...,v a-1,m,vsl (t+1)), define i′∈[an,a-1], where v i′,m,vsl (t+1) represents the i′th variable speed limit control cell of any individual in any generation of population, representing the variable speed limit value of the i′th cell, and the ramp adjustment rate sub-gene of any individual in any generation of population is recorded as r(t+1);

[0082] Step b: predict the sending and receiving capabilities of each cell in the t+1 time period;

[0083] Step b1: predict the sending capacity δ of the i′th variable speed limit control implementation cell in the t+1 time period according to the formula i′,m,vsl (t+1) and receiving capacity σ i′,m,vsl (t+1);

[0084] σ i′,m,vsl (t+1)=min{v i′,m (t+1)·k i′,m (t+1)·n i′,m ,v i′,m,vsl (t+1)·k i′,m,vsl (t+1)·n i′,m}i′∈[an,a-1] (19)

[0085] δ i′,m,vsl (t+1)={w m ·(k jam,m -k i′,m (t+1))·n i′,m ,v i′,m,vsl (t+1)·k i′,m,vsl (t+1)·n i′,m}i′∈[an,a-1] (20)

[0086] In formulas (19) to (20), v i′,m,vsl (t+1), k i′,m,vsl (t+1) represents the speed and the corresponding average density of the i′th variable speed control cell in the t+1 time period, and is obtained by formula (21); v i′,m (t+1), k i′,m (t+1) represents the average speed and average density of the i′th variable speed control cell in the t+1 time period, and is calculated by equations (7) to (10);

[0087]

[0088] Calculate the sending and receiving capabilities of other mainline cells according to formula (1) to formula (2);

[0089] Step b2: Predict the sending capacity σ of the ramp control implementation cell in the t+1 time period according to formula (22): J,r,ratio (t+1);

[0090] σ J,r,ratio (t+1)=min{v j,r (t+1)·k j,r (t+1)·n j,r ,r(t+1)·Q r ·n j,r}·Δt(22)

[0091] In formula (22), r(t+1) represents the ramp adjustment rate corresponding to the t+1 time period;

[0092] The receiving capacity δ of the ramp control implementation cell is calculated according to formula (4): J,r (t);

[0093] Calculate the sending and receiving capabilities of other ramp cells according to formulas (3) and (4);

[0094] Step c: Use formula (23) to determine the transmission flow q of the i-th mainline cell in the t+1 time period i,m (t+1);

[0095]

[0096] When i=a, if δ i+1,m (t+1)<σ i,m (t+1)+σ J,r,ratio (t+1), then using formula (24) we can get q i,m (t+1), otherwise, let q i,m (t+1)=σi,m (t+1);

[0097] q i,m (t+1)=max{δ i+1,m (t+1)-σ J,r,ratio (t+1),δ i+1,m (t+1)(1-p on (t+1))} (24)

[0098] In formula (24), σ J,r,ratio (t+1) represents the sending capacity of the ramp control implementation cell in the t+1 time period, p on (t+1) represents the confluence ratio between the mainline area and the ramp area in the t+1th time period, and p on (t+1)=p on (t);

[0099] Calculate the transmission flow of other main lines and ramp cells in the t+1 time period according to formula (5) to formula (6);

[0100] Step d: Determine the traffic volume q of the ramp control implementation cell under the variable speed limit in the t+1 time period J,r (t+1):

[0101]

[0102] Step e: Predict the average density k of the i-th main line cell in the t+2 time period according to equations (26) and (27). i,m (t+2) and the average density k of the j-th ramp cell j,r (t+2);

[0103]

[0104] k j,r (t+2)=k j,r (t+1)+(q j-1,r (t+1)-q j,r (t+1))×Δt / l j,r (27)

[0105] Step f: Predict the average velocity v of the i-th main line cell in the t+2 time period according to equations (28) to (29). i,m (t+2) and the average velocity v of j ramp cells j,r (t+2);

[0106]

[0107]

[0108] An electronic device according to the present invention includes a memory and a processor, and is characterized in that the memory is used to store a program that supports the processor to execute the dual-layer collaborative control method, and the processor is configured to execute the program stored in the memory.

[0109] The present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, which is characterized in that when the computer program is run by a processor, the steps of the two-layer collaborative control method are executed.

[0110] Compared with the existing technology, the beneficial technical effects of the present invention are embodied in:

[0111] 1. In an intelligent connected environment, the present invention provides a two-layer collaborative control method that takes into account traffic risks in expressway merging areas. Compared with the traditional static rules or single-layer optimization strategies, a two-layer planning model is established. The upper layer is variable speed limit control, and the lower layer is refined ramp control. This achieves differentiated control of different road sections, which is more refined and targeted than the previous single-layer control.

[0112] 2. Compared with existing technologies, the bi-level programming model established in this invention has multiple objectives at each level, allowing for step-by-step optimization. This allows for a clearer decomposition of the optimization control problem for the mainline and ramps, with different traffic efficiency indicators as control objectives. Separately modeling the two sub-problems, the mainline and ramp, reduces problem complexity and improves solution efficiency. Compared to most multi-objective planning methods, bi-level programming control is less prone to falling into local optimality, facilitating global optimization. Bi-level collaborative decision-making enables effective information sharing and intelligent optimization between the mainline and ramps, significantly improving traffic control efficiency.

[0113] 3. Compared with existing technologies, this invention comprehensively considers multiple traffic risk optimization models for collision risk and congestion evaluation, and comprehensively assesses traffic conditions. The upper and lower layer collaborative control strategies comprehensively optimize traffic flow distribution, effectively reducing collision risks. Dynamic variable speed limit control is mainly aimed at reducing collision risks, while ramp signal control focuses on alleviating congestion. Compared with existing technologies, the collaborative optimization of the two-layer control strategy better takes into account and balances collision risk control and congestion improvement, significantly improving the safety and traffic efficiency of complex intersections.

[0114] 4. Compared with existing technologies, the present invention comprehensively considers traffic efficiency indicators such as maximum traffic turnover in the merging area, minimum collision risk, minimum average ramp waiting time, and minimum ramp vehicle energy consumption. This avoids the situation where coordinated control reduces the collision risk in the merging area while significantly reducing traffic turnover, and avoids the frequent starting and stopping of vehicles caused by the control target optimization process, thereby improving traffic stability and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 It is the overall flow chart of the present invention;

[0116] Figure 2 is a flow chart of the decision-making method of the present invention;

[0117] Figure 3 This is a schematic diagram of the expressway merging area of ​​the present invention; DETAILED DESCRIPTION

[0118] In this embodiment, a two-layer cooperative control method for considering traffic risks in expressway merging areas in a networked environment refers to the intersection area of ​​the main line area and the ramp area of ​​the expressway; the vehicle driving direction is taken as the positive direction, and the length of the merging area is taken as the standard length. The upstream area of ​​the merging area is divided into a main line cells according to the standard length, the merging area is used as a merging cell and recorded as the a+1th main line cell, and the downstream area of ​​the merging area is divided into b main line cells according to the standard length, thereby obtaining a+b+1 mainline cells; any one of the a+b+1 mainline cells is recorded as the i-th mainline cell according to the vehicle's travel direction, the ramp area is divided into J ramp cells according to the standard length, and the ramp cell adjacent to the a+1-th mainline cell is taken as the J-th ramp cell, and any ramp cell is defined as the j-th ramp cell; the n mainline cells upstream of the a-th mainline cell are set as variable speed limit control implementation cells, and the J-th ramp cell is set as the ramp control implementation cell, 1≤n≤a-1; specifically, if Figure 1 As shown, the method includes the following steps:

[0119] Step 1: Use roadside intelligent devices to obtain traffic flow parameters of each cell on the expressway mainline and ramp in time period t, and use them to predict traffic flow parameters in time period t+1;

[0120] Step 1.1: Calculate the sending and receiving capabilities of each cell in time period t:

[0121] According to formula (1) to formula (2), the sending capacity σ of the i-th main line cell in the t-th time period is calculated i,m (t) and receiving capacity δ i,m (t):

[0122] σ i,m (t) = min{v i,m (t)·k i,m (t)·n i,m ,Q m ·n i,m}·Δt (1)

[0123] δ i,m (t) = {w m ·(k jam,m -ki,m (t))·n i,m ,Q m ·n i,m}·Δt (2)

[0124] In formula (1) to formula (2), v i,m (t), k i,m (t) represents the average velocity and average density of the i-th main line cell in the t-th time period; k jam,m , w m , Q m , n i,m They represent the mainline blocking density, wave speed, single lane maximum capacity and number of lanes of the i-th mainline cell respectively, and Δt is the interval between adjacent time periods;

[0125] According to formula (3) to formula (4), the sending capacity σ of the j-th ramp cell in the t-th time period is calculated j,r (t) and receiving capacity δ j,r (t);

[0126] σ j,r (t) = min{v j,r (t)·k j,r (t)·n j,r ,Q r ·n j,r}·Δt (3)

[0127] δ j,r (t) = {w r ·(k jam,r -k j,r (t))·n j,r ,Q r ·n j,r}·Δt (4)

[0128] In formula (3) to formula (4), v j,r (t), k j,r (t) represents the average speed and average density of the j-th ramp cell in the t-th time period, k jam,r , w r , Q r , n j,r They represent the ramp blocking density, wave speed, maximum single lane capacity and number of lanes of the j-th ramp cell respectively;

[0129] Step 1.2: Calculate the transmission flow q of the i-th mainline cell in the t-th time period according to formula (5): i,m (t);

[0130]

[0131] In formula (5), q i,m (t) represents the traffic volume transmitted downstream by the i-th mainline cell in the t-th time period, p on (t) represents the confluence ratio between the mainline area and the ramp area in the tth time period; σ a,m (t) represents the sending capacity of the ath mainline cell in the tth time period, δ a+1,m (t) represents the receiving capacity of the a+1th mainline cell in the tth time period, σ J,r (t) represents the sending capacity of the J-th ramp cell in the t-th time period;

[0132] Step 1.3: Predict the transmission flow q of the jth ramp cell in the tth time period according to formula (6): j,r (t);

[0133]

[0134] In formula (6), σ J,r (t) represents the sending capacity of the J-th ramp cell in the t-th time period; δ j+1,r (t) represents the receiving capacity of the j+1th ramp cell in the tth time period;

[0135] Step 1.4: Predict the average density k of the i-th main line cell in the t+1 time period according to equations (7) and (8). i,m (t+1) and the average density k of the j-th ramp cell j,r (t+1):

[0136]

[0137] k j,r (t+1)=k j,r (t)+(q j-1,r (t)-q j,r (t))×Δt / l j,r (8)

[0138] In formula (7) to formula (8), l i,m represents the length of the i-th main line cell, l j,r represents the length of the j-th ramp cell;

[0139] Step 1.5: Predict the average velocity v of the i-th main line cell in the t+1 time period according to equations (9) and (10): i,m (t+1) and the average velocity v of the j-th ramp cell j,r (t+1);

[0140]

[0141]

[0142] In formula (9) to formula (10), V f,m represents the free flow velocity of the main line cell, w m represents the congestion wave propagation speed of the main line cell, k cm represents the critical density of the main line cell, V f,r represents the free flow velocity of the ramp cell, w r represents the congestion wave propagation speed of the ramp cell, k cr represents the critical density of ramp cells.

[0143] Step 2: Figure 2 As shown in the figure, the traffic flow parameters of the t+1 time period are input into the accident risk assessment model and the congestion risk assessment model respectively, and the accident risk value CR of the merging cell in the t+1 time period is calculated. a+1,m (t+1) and congestion risk value JAM a+1,m (t+1); Input the traffic flow parameters of the t+1 time period into the ramp queue length model and calculate the queue length W(t+1) of ramp cell J in the t+1 time period;

[0144] In step 2, the accident risk assessment model is constructed using formula (11), the congestion risk assessment model is constructed using formula (12), and the ramp queue length model is constructed using formula (13):

[0145]

[0146]

[0147] W(t+1)=W(t)+(q J,r (t)-q J-1,r (t))·Δt (13)

[0148] In formula (11)-(13), x1, x2, x3, x4, x5 are the coefficients of five variables, β is the intercept, and v a-3,m (t+1) represents the average velocity of the a-3th main line cell in the t+1th time period, k a-2,m (t+1) represents the average density of the a-2th main line cell in the t+1th time period, v a+1,m (t+1) represents the average velocity of the a+1th main line cell in the t+1th time period, q a+2,m (t) represents the transmission flow of the a+2th mainline cell in the tth time period, dq a-3,a+2,m V represents the difference in transmission flow between the a-3th mainline cell and the a+2th mainline cell in the tth time period, max Indicates the maximum speed limit on the expressway, v a+1,m(t+1) represents the average velocity of the a+1th main line cell in the t+1th time period, q a+1,m (t) represents the transmission flow of the a+1th mainline cell in the tth time period, Q m represents the maximum traffic capacity of a single lane, q J-1,r (t) represents the transmission flow of the J-1th ramp cell upstream of ramp cell J in the tth time period, W(t) represents the number of vehicles queuing in ramp cell J in the tth time period. When t = 1, the number of vehicles queuing in the Jth ramp cell in the tth time period, W(t), is detected by the roadside intelligent device. Δt is the interval between adjacent time periods.

[0149] Step 3: Determine CR a+1,m (t+1) Whether the accident risk threshold Δ is exceeded CR or JAM a+1,m (t+1) Whether the congestion risk threshold Δ is exceeded JAM If yes, go to step 4, otherwise go to step 10;

[0150] Step 4: Based on the various control objectives of the main line and the ramp, a two-level planning model for the merging area in the t+1 time period is constructed. This model includes an upper-level planning model with the control objectives of minimizing the accident risk and maximizing the traffic turnover in the merging area, and a lower-level planning model with the control objectives of minimizing the average waiting time of ramp vehicles and the energy consumption of ramp vehicles. The upper-level planning model is constructed using Equation (14), and the lower-level planning model is constructed using Equation (15):

[0151]

[0152] In formula (14)-formula (15), J U (t+1) represents the objective function value of the upper-level planning model in the t+1 time period, J D (t+1) represents the objective function value of the lower-level planning model in the t+1th time period, μ1, μ2, μ3, α1 and α2 are five-dimensional balance coefficients, l a+1,m represents the length of the a+1th main line cell, C D is the drag coefficient, A is the projected area in the direction of the car's travel, ρ is the air density, Z is the car's gravity, f is the rolling resistance coefficient, and v J,r (t+2) represents the average speed of the ramp control cell in the t+2 time period;

[0153] Formula (16) is used to construct the constraints of the two-level planning model of the merging area:

[0154]

[0155] In formula (16), v i(t) represents the variable speed limit value of the i-th variable speed limit control cell in the t-th time period, v i-1 (t) represents the variable speed limit value of the i-1th variable speed limit control cell in the tth time period, v i (t+1) represents the variable speed limit value of the i-th variable speed limit control cell in the t+1 time period, spd diff,s Indicates the speed limit difference threshold of adjacent variable speed limit control cells in the same time period, spd diff,t represents the speed limit difference threshold of the same variable speed limit control cell in adjacent time periods, v min Indicates the minimum speed limit on the expressway, v max Indicates the maximum speed limit on the expressway.

[0156] Step 5: Determine whether the ramp queue length W(t+1) exceeds the threshold Δ W , if yes, then execute the coordinated control strategy of step 6, otherwise, execute the ramp control strategy of step 7;

[0157] Step 6: Use a hybrid genetic simulated annealing algorithm to iteratively solve the merging area bi-level programming model to determine the optimal variable speed limit and optimal ramp adjustment rate in the t+1 time period, and then execute step 8;

[0158] The hybrid genetic simulated annealing algorithm is solved in the following steps:

[0159] Step 6.1, define the population size as M, the current number of evolutions as G, and the maximum number of evolution generations as G max , the population crossover probability is p c , the population mutation probability is p v , initialize G = 1;

[0160] Step 6.2: Generate the G-generation genetic population in the t+1 time period

[0161] Step 6.2.1: Generate M chromosomes in the t+1 time period according to formula (16) and form the G-th generation variable speed-limited population. in, Represents the G-th generation variable speed limit population in the t+1 time period The sth chromosome in , and each chromosome represents a variable speed limit control strategy, Represents the sth variable speed limit control strategy The variable speed limit value of the e-th variable speed limit control cell in ; e=1,2,...,n,s=1,2,...,M;

[0162] Step 6.2.2: Randomly generate the G-th generation variable speed limit population Chromosome s The ramp regulation rate real number sub-gene string in, express The corresponding s-th ramp regulation rate sub-gene in , and Thus, the G generation genetic population is obtained in, represents the sth individual, s=1,2,...,M;

[0163] Step 6.3: Obtain the G-generation genetic population The sth individual The corresponding minimum fitness value Ramp regulatory subgene Thus, the sth chromosome and Form the sth new individual, thus reaching the Gth generation genetic population The Gth generation new genetic population is composed of the chromosome of each individual and the ramp adjustment rate sub-gene of its minimum fitness value

[0164] Step 6.3.1, the G generation genetic population The sth individual Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the traffic flow state quantity predicted by each cell of the Gth generation in the t+2th time period;

[0165] Step 6.3.2: Substitute the traffic flow state predicted by each cell in the t+2th time period into the lower-level planning model, and use the inverse of the lower-level planning model as the fitness function to obtain the Gth generation genetic population. The sth individual The fitness set in, express The corresponding s-th ramp regulation rate sub-gene The fitness value of

[0166] Step 6.3.3: Genetic population of generation G The sth individual The fitness value set Compare the fitness values ​​of each ramp regulation rate sub-gene in the and select the minimum fitness value The corresponding ramp regulation rate sub-gene is recorded as

[0167] Step 6.4, G generation genetic population The chromosomes of each individual and the ramp regulation rate sub-genes of each individual are crossover and mutation operations are performed to generate the offspring population of the G generation population. For offspring population Perform selection operation to obtain the genetic population of the G+1 generation

[0168] Step 6.5: Calculate the genetic population of the G+1 generation according to the process of step 6.3 The minimum fitness value corresponding to all individuals in in, express The minimum fitness value corresponding to the Mth individual in The sth chromosome in and The corresponding ramp regulation rate sub-gene Forming the G+1 generation new genetic population The sth individual;

[0169] like Less than Then retain the G+1 generation new genetic population The sth individual Otherwise, Assign to Will Assign to the sth new individual Thus updating the G+1 generation new genetic population The sth new individual;

[0170] Step 6.6, assign G+1 to G, and judge whether G<G max Is it true? If so, go to step 6.4; otherwise, it means that G is completed. max The Gth iteration is obtained max New genetic population s=1,2,...,M;

[0171] Step 6.7, Performing an annealing operation;

[0172] Step 6.7.1, define the initial temperature as H0 and the final temperature as H end , temperature drop coefficient α and maximum number of cycles Inter; define the current temperature as H k , and initialize H k =H0; the current number of cycles is inter, and initialize inter=1; As the current inter generation population, denoted as

[0173] Step 6.7.2, Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the traffic flow state quantity predicted by each inter-generation cell in the t+2 time period. The inverse of formula (14) is used as the fitness function to obtain The upper fitness value of each individual in the inter generation is selected, and the individual corresponding to the minimum fitness value is recorded as the optimal individual of the inter generation. in, express Minimum fitness value The corresponding chromosomes express The corresponding ramp regulation rate sub-gene;

[0174] Step 6.7.3: According to the Metropolis criterion and H k ,right Each individual is selected to obtain the inter+1 generation population

[0175] Step 6.7.4: Determine whether inter = Inter. If so, proceed to step 6.7.5. Otherwise, assign inter + 1 to inter and proceed to step 6.7.2.

[0176] Step 6.7.5, if H k =H end , which means the optimal variable speed limit value and the optimal ramp adjustment rate in the t+1 time period are obtained Otherwise, αH k Assign to H k , proceed to step 6.7.3.

[0177] Step 7: Use a genetic algorithm to iteratively solve the upper-level planning model in the merging area bi-level planning model to determine the optimal ramp adjustment rate in the t+1 time period, and then execute step 9;

[0178] Step 7.1. Initialize the population size M, the current number of evolutions G, and the maximum number of evolution generations G max , population crossover probability p c , population mutation probability p v , initialize G = 1;

[0179] Step 7.2: Randomly generate M chromosomes in the t+1 time period. Each chromosome represents a ramp adjustment rate. The G-th iteration set is obtained as And r s (t+1)∈[0,1], s=1,2,...,M;

[0180] Step 7.3: Based on the improved cellular transmission model, predict the traffic flow state of each cell in the t+2 time period under the ramp adjustment rate corresponding to the sth chromosome in the t+1 time period;

[0181] Step 7.4: Substitute the traffic flow state quantity of the t+1th time period corresponding to the sth chromosome in the Gth generation population into the upper planning model, and use the inverse of the upper planning model formula (14) as the fitness function to obtain the fitness set of the Gth generation population. It represents the corresponding fitness value of the sth chromosome of the Gth generation under the objective function in the t+1th time period;

[0182] Step 7.5: Set the fitness values ​​of all chromosomes for the Gth time Compare the fitness values ​​of each chromosome in the G generation and save the chromosome with the minimum fitness value of each group This value indicates that this strategy is to minimize the traffic risk value of the merging area in the t+2 time period. The minimum fitness value is recorded as

[0183] Step 7.6: Set the Gth generation population to a population size of M Perform crossover and mutation to generate the G generation population The offspring population is Its population size is also M, and the offspring population Perform selection operations to keep the population size M, and the resulting chromosomes serve as the parent population of the G+1 generation. s=1,2,...,M;

[0184] Step 7.7: Calculate the parent population of the G+1 generation according to steps 7.3 to 7.4. The fitness value set of all chromosomes in The minimum fitness value of each group of chromosomes in the G+1 generation is The corresponding chromosome is denoted as like Less than in step 7.5 Then the updated minimum fitness value is And save its corresponding chromosome; otherwise, use the chromosome corresponding to the minimum fitness value in the G generation chromosome Replace the chromosome with the minimum fitness value in the G+1 generation chromosome and enter the next generation;

[0185] Step 7.8: Assign G+1 to G and determine whether G < G max Is it true? If so, go to step 7.6; otherwise, it means that G is completed. max The final population set is recorded as The population in The gene is the optimal solution, which is the optimal ramp adjustment ratio ratio(t+1);

[0186] In specific implementation, the improved cellular transmission model predicts the traffic flow state of any generation in the t+2 time period according to the following steps:

[0187] Step a: record the chromosome of any individual in any generation of population as (v a-n,m,vsl (t+1),...,v a-1,m,vsl (t+1)), define i′∈[an,a-1], where v i′,m,vsl (t+1) represents the i′th variable speed limit control cell of any individual in any generation of population, representing the variable speed limit value of the i′th cell, and the ramp adjustment rate sub-gene of any individual in any generation of population is recorded as r(t+1);

[0188] Step b: predict the sending and receiving capabilities of each cell in the t+1 time period;

[0189] Step b1: predict the sending capacity δ of the i′th variable speed limit control implementation cell in the t+1 time period according to the formula i′,m,vsl (t+1) and receiving capacity σ i′,m,vsl (t+1);

[0190] σ i′,m,vsl (t+1)=min{v i′,m (t+1)·k i′,m (t+1)·n i′,m ,v i′,m,vsl (t+1)·k i′,m,vsl (t+1)·n i′,m}i′∈[an,a-1] (19)

[0191] δ i′,m,vsl (t+1)={w m ·(k jam,m -k i′,m (t+1))·n i′,m ,v i′,m,vsl (t+1)·k i′,m,vsl (t+1)·n i′,m}i′∈[an,a-1] (20)

[0192] In formulas (19) to (20), v i′,m,vsl (t+1), k i′,m,vsl (t+1) represents the speed and the corresponding average density of the i′th variable speed limit control cell in the t+1 time period, and is obtained by formula (21); v i′,m (t+1), ki′,m (t+1) represents the average speed and average density of the i′th variable speed control cell in the t+1 time period, and is calculated by equations (7) to (10);

[0193]

[0194] Calculate the sending and receiving capabilities of other mainline cells according to formula (1) to formula (2);

[0195] Step b2: Predict the sending capacity σ of the ramp control implementation cell in the t+1 time period according to formula (22): J,r,ratio (t+1);

[0196] σ J,r,ratio (t+1)=min{v j,r (t+1)·k j,r (t+1)·n j,r ,r(t+1)·Q r ·n j,r}·Δt(22)

[0197] In formula (22), r(t+1) represents the ramp adjustment rate corresponding to the t+1 time period;

[0198] The receiving capacity δ of the ramp control implementation cell is calculated according to formula (4): J,r (t);

[0199] Calculate the sending and receiving capabilities of other ramp cells according to formulas (3) and (4);

[0200] Step c: Use formula (23) to determine the transmission flow q of the i-th mainline cell in the t+1 time period i,m (t+1);

[0201]

[0202] When i=a, if δ i+1,m (t+1)<σ i,m (t+1)+σ J,r,ratio (t+1), then using formula (24) we can get q i,m (t+1), otherwise, let q i,m (t+1)=σ i,m (t+1);

[0203] q i,m (t+1)=max{δ i+1,m (t+1)-σ J,r,ratio (t+1),δ i+1,m (t+1)(1-p on (t+1))}(24)

[0204] In formula (24), σ J,r,ratio (t+1) represents the sending capacity of the ramp control implementation cell in the t+1 time period, p on (t+1) represents the confluence ratio between the mainline area and the ramp area in the t+1th time period, and p on (t+1)=p on (t);

[0205] Calculate the transmission flow of other main lines and ramp cells in the t+1 time period according to formula (5) to formula (6);

[0206] Step d: Determine the traffic volume q of the ramp control implementation cell under the variable speed limit in the t+1 time period J,r (t+1):

[0207]

[0208] Step e: Predict the average density k of the i-th main line cell in the t+2 time period according to equations (26) and (27). i,m (t+2) and the average density k of the j-th ramp cell j,r (t+2);

[0209]

[0210] k j,r (t+2)=k j,r (t+1)+(q j-1,r (t+1)-q j,r (t+1))×Δt / l j,r (27)

[0211] Step f: Predict the average velocity v of the i-th main line cell in the t+2 time period according to equations (28) to (29). i,m (t+2) and the average velocity v of j ramp cells j,r (t+2);

[0212]

[0213]

[0214] Step 8: In the t+1 time period, the variable speed limit control implementation cell implements coordinated control according to the optimal variable speed limit value, and the ramp control cell implements coordinated control according to the optimal ramp adjustment rate, and then executes step 10;

[0215] Step 9: In the t+1 time period, the ramp control cell performs ramp control according to the optimal ramp regulation rate, and executes step 10.

[0216] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above-mentioned two-layer collaborative control method, and the processor is configured to execute the program stored in the memory.

[0217] In this embodiment, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned two-layer collaborative control method are executed.

[0218] like Figure 3 As shown, in this embodiment, the number of main line lanes n i,m =2, number of ramp lanes n j,r =1 as an example, taking the vehicle driving direction as the positive direction, taking the road upstream of the expressway merging area as being divided into 6 consecutive cells as an example, and numbering them in sequence according to the driving direction, wherein any area is numbered i, i = 1, 2, 3, 4, 5, 6, wherein cells 2 and 3 are mainline variable speed limit control implementation cells, and cell 5 is the merging cell, taking the road upstream of the entrance ramp corresponding to the merging area as being divided into 3 consecutive sections as an example, each section is a cell, and numbering them in sequence according to the driving direction, wherein any area is numbered j, j = 1, 2, 3, and the ramp cell connected to the merging cell 5 is numbered J = 3, and the basic graph model parameters of the road section are set, and the long interval of each time period is set to Δt = 1 min.

[0219] Get the basic traffic flow diagram parameters of the merging area, the main line free flow speed v f,m =100 (km / h), single lane capacity q m =1500 (veh / h), blocking density k jam,m =120 (veh / km), critical density k cm =15(veh / km), wave speed w m =14.29 (km / h); ramp free flow speed v f,r =50 (km / h), single lane capacity q r =1000(veh / h), blocking density k jam,r =100(veh / km), critical density k cr =20(veh / km), wave speed w r =12.50 (km / h), with risk threshold Δ CR =0.5,Δ JAM =0.5, the ramp queue length threshold is Δ W =10 as an example;

[0220] The current traffic density and average speed of each cell on the main line detected by intelligent roadside equipment are Ki,m(t)={24, 26.7110, 41.5367, 64.4118, 21.0746, 21.0656}, V i,m (t)={100,88,75,39,100,100}, the current traffic density and average speed of each section of the ramp are Kj, r(t)={18,17.9931,18.9525}, V j,r (t) = {50, 50, 50}; detect the traffic volume q of the upstream input starting cell in the next period 0,m =1596(veh / h),q 0,r =608 (veh / h);

[0221] According to step 1, the traffic flow parameters for the next time period are predicted, and the sending and receiving capacities of each cell on the main line σ i,m (t)={2400,2167,2925,3000,2107,2106},δ i,m = {3000,3000,2843,2176,3000,3000} The sending and receiving capabilities of each cell on the ramp σ j,r (t) = {899, 899, 947}, δ j,r ={1000,1000,1000}; the traffic volume transmitted by each cell of the main line to its downstream cell is q i,m (t) = {2400, 2671, 2176, 2111, 2107, 2106}, with a density of K i,m (t+1)={24,24.2008,46.1168,65.0201,21.1050,21.0739}, the average speed is V i,m (t+1)={100,85,75,38.7950,100,100}; the traffic volume transmitted from each ramp cell to the downstream cell is q j,r (t) = {899,899,889}; average density is K j,r (t+1)={18,18,19.0520}, the average speed is V j,r (t+1) = {50,50,50}; the risk value of the t+1 period is CR 5,m (t+1)=0.71, the congestion risk value is JAM 5,m (t+1)=0.76JAM(t+1)=0.76, the ramp queue length is W(t+1)=12, at this time the risk value exceeds the threshold, the ramp queue length exceeds the threshold, the coordinated control measures are enabled, and the constraint condition (16) is set as According to step 6, a population of variable speed limits for the main line is randomly generated, with a population size of M = 100, forming a set

[0222] Generate M=100 ramp adjustment rates for each speed limit population in the variable speed limit population to form a combined population According to step 6, the traffic flow state corresponding to each combination population is predicted. According to the fitness function with the inverse of the lower control target, the optimal ramp adjustment rate corresponding to each speed limit population is obtained by genetic algorithm. Iteration is performed until the optimal result is obtained or the maximum number of iterations is reached, thereby obtaining the coordinated control population. The inverse of the upper control target is used as the fitness function to calculate the fitness value of the population of the collaborative control solution value set. After the population is selected, mutated, and crossover, the traffic state variables of the next generation population are predicted, the corresponding fitness value is calculated, and the simulated annealing operation is performed to determine whether a better new individual is generated until the optimal solution is found or the specified number of iterations is reached. The optimal variable speed limit value and the optimal ramp adjustment rate that minimize the risk value in the current time period are obtained [v vsl1 (t+1)=55,v vsl2 (t+1)=70,,ratio(t+1)=0.73], start the coordinated control of the t+1 time period according to this value; after assigning t+1 to t, if t reaches the time threshold T, stop the control; otherwise, return to step 1 and execute sequentially.

[0223] In this embodiment, the method concept of the present invention is not limited to the coordinated control of traffic risks from the expressway merging area to the upstream of the main line and the ramp area. Other embodiments obtained by ordinary technicians in this field without creative changes are within the scope of protection of the present invention.

Claims

1. A two-tiered collaborative control method for considering traffic risks in expressway merging areas in a connected environment. The expressway merging area is the intersection of the mainline area and the ramp area of ​​the expressway. Taking the vehicle travel direction as the positive direction and the length of the intersection area as the standard length, the upstream area of ​​the intersection area is divided into a mainline cells according to the standard length, the intersection area is used as a merging cell and recorded as the a+1th mainline cell, and the downstream area of ​​the intersection area is divided into b mainline cells according to the standard length, thereby obtaining a+b+1 mainline cells; according to the vehicle travel direction, any mainline cell among the a+b+1 mainline cells is recorded as the i-th mainline cell, the ramp area is divided into J ramp cells according to the standard length, and the ramp cell adjacent to the a+1th mainline cell is used as the J-th ramp cell, and any ramp cell is defined as the j-th ramp cell; the n mainline cells upstream of the a-th mainline cell are set as variable speed limit control implementation cells, and the J-th ramp cell is set as the ramp control implementation cell. ; The method is carried out as follows: Step 1: Use roadside intelligent devices to obtain traffic flow parameters of each cell on the expressway mainline and ramp in time period t, and use them to predict traffic flow parameters in time period t+1; Step 2: Input the traffic flow parameters of the t+1 time period into the accident risk assessment model and the congestion risk assessment model respectively, and calculate the accident risk value of the merging cell in the t+1 time period. and congestion risk value ; Input the traffic flow parameters of the t+1 time period into the ramp queue length model, and calculate the The queue length of ramp cell J in the time period ; Step 3: Judgment Whether the accident risk threshold is exceeded or Whether the congestion risk threshold is exceeded If yes, go to step 4, otherwise go to step 10; Step 4: Based on the various control objectives of the main line and ramp, construct the A two-level planning model for merging areas under time periods, including an upper-level planning model with the control objectives of minimizing the accident risk and maximizing the traffic turnover in the merging area, and a lower-level planning model with the control objectives of minimizing the average waiting time and energy consumption of ramp vehicles; Step 5: Determine the length of the ramp queue Whether it exceeds the threshold , if yes, then execute the cooperative control strategy of step 6, otherwise, execute the ramp control strategy of step 7; Step 6: Use the hybrid genetic simulated annealing algorithm to iteratively solve the confluence area bi-level programming model to determine the The optimal variable speed limit value and the optimal ramp adjustment rate in the time period, and executing step 8; Step 7: Use genetic algorithm to iteratively solve the upper level planning model in the two-level planning model of the confluence area to determine the The optimal ramp regulation rate for the time period and executing step 9; Step 8: In the t+1 time period, the variable speed limit control implementation cell implements coordinated control according to the optimal variable speed limit value, and the ramp control cell implements coordinated control according to the optimal ramp adjustment rate, and then executes step 10; Step 9: In the t+1 time period, the ramp control cell performs ramp control according to the optimal ramp regulation rate, and executes step 10; Step 10: Assign to Afterwards, if Arrival time threshold , then stop the control, otherwise, return to step 1 and execute sequentially.

2. The dual-layer collaborative control method according to claim 1, characterized in that: The step 1 comprises: Step 1.1: Calculate the sending and receiving capabilities of each cell in time period t: According to formula (1) to formula (2), the sending capacity of the i-th main line cell in the t-th time period is calculated and receiving capabilities : (1) (2) In formula (1) to formula (2), , Respectively represent The average speed and average density of the i-th main line cell in the time period; , , , They represent the mainline blocking density, wave speed, maximum single lane capacity and the number of lanes of the i-th mainline cell, respectively. is the interval between adjacent time periods; According to formula (3) to formula (4), the sending capacity of the j-th ramp cell in the t-th time period is calculated and receiving capabilities ; (3) (4) In formula (3) to formula (4), , Respectively represent The average speed and average density of the j-th ramp cell in the time period, , , , They represent the ramp blocking density, wave speed, maximum single lane capacity and the number of lanes of the jth ramp cell respectively; Step 1.2: Calculate the The transmission flow of the i-th main line cell in the time period ; (5) In formula (5), Indicates the The traffic volume transmitted downstream by the i-th mainline cell in the time period, Indicates the The confluence ratio between the mainline area and the ramp area in each time period; represents the sending capacity of the ath mainline cell in the tth time period, represents the receiving capacity of the a+1th mainline cell in the tth time period, represents the sending capacity of the J-th ramp cell in the t-th time period; represents the sending capacity of the ith mainline cell in the tth time period, represents the receiving capacity of the i+1th mainline cell in the tth time period; Step 1.3: Predict the The transmission flow of the j-th ramp cell in the time period ; (6) In formula (6), represents the sending capacity of the J-th ramp cell in the t-th time period; represents the receiving capacity of the j+1th ramp cell in the tth time period; represents the sending capacity of the j-th ramp cell in the t-th time period, represents the receiving capacity of the j+1th ramp cell in the tth time period; Step 1.4: Predict the first The average density of the i-th main line cell in the time period and the average density of the j-th ramp cell : (7) (8) In formula (7) to formula (8), represents the length of the i-th main line cell, represents the length of the j-th ramp cell; Step 1.5: Predict the first The average speed of the i-th main line cell in the time period and the average speed of the j-th ramp cell ; (9) (10) In formula (9) to formula (10), represents the free flow velocity of the mainline cell, represents the congestion wave propagation speed of the main line cell, represents the critical density of the main line cell, represents the free flow velocity of the ramp cell, represents the congestion wave propagation speed of the ramp cell, represents the critical density of ramp cells.

3. The dual-layer cooperative control method according to claim 2, characterized in that: In step 2, the accident risk assessment model is constructed using formula (11), the congestion risk assessment model is constructed using formula (12), and the ramp queue length model is constructed using formula (13): (11) (12) (13) In formula (11) to formula (13), are the coefficients of 5 variables, is the intercept, represents the average velocity of the a-3th main line cell in the t+1th time period, represents the average density of the a-2th main line cell in the t+1th time period, represents the average velocity of the a+1th main line cell in the t+1th time period, represents the transmission flow of the a+2th mainline cell in the tth time period, represents the difference in transmission flow between the a-3th mainline cell and the a+2th mainline cell in the tth time period, Indicates the maximum speed limit on the expressway. represents the transmission flow of the a+1th mainline cell in the tth time period, represents the transmission flow of the J-1th ramp cell upstream of ramp cell J in the tth time period, represents the number of queued vehicles of ramp cell J in the t-th time period. When t=1, the number of queued vehicles of ramp cell J in the t-th time period is detected by roadside intelligent equipment. , The interval between adjacent time periods.

4. The dual-layer collaborative control method according to claim 3 is characterized in that: In step 4, the upper-level planning model is constructed using formula (14) and the lower-level planning model is constructed using formula (15): (14) (15) In formula (14)-formula (15), represents the objective function value of the upper-level planning model in the t+1th time period, represents the objective function value of the lower-level planning model in the t+1th time period, 、 、 、 and is the 5-dimensional balance coefficient, represents the length of the a+1th main line cell, is the drag coefficient, Indicates the projected area in the direction of the car's travel, represents the air density, Z represents the vehicle's gravity, represents the rolling resistance coefficient, represents the average speed of the ramp control cell in the t+2 time period; Formula (16) is used to construct the constraints of the two-level planning model of the merging area: (16) In formula (16), Indicates the first The variable speed limit value of the variable speed limit control cell, Indicates the first The variable speed limit value of the variable speed limit control cell, Indicates the first The variable speed limit value of the variable speed limit control cell, Indicates the speed limit difference threshold of adjacent variable speed limit control cells in the same time period, represents the speed limit difference threshold of the same variable speed limit control cell in adjacent time periods, Indicates the minimum speed limit on the expressway. Indicates the maximum speed limit on the expressway.

5. The dual-layer cooperative control method according to claim 4, characterized in that: The hybrid genetic simulated annealing algorithm in step 6 is solved according to the following steps: Step 6.1, define the population size as M, the current number of evolutions as , the maximum evolutionary generation is , the population crossover probability is , the population mutation probability is ,initialization ; Step 6.2: Generate the G-generation genetic population in the t+1 time period ; Step 6.2.1: Generate M chromosomes in the t+1 time period according to formula (16) and form the G-th generation variable speed-limited population. ,in, Represents the G-th generation variable speed limit population in the t+1 time period The sth chromosome in , and each chromosome represents a variable speed limit control strategy, , Represents the sth variable speed limit control strategy Middle The variable speed limit value of a variable speed limit control cell; , ; Step 6.2.2: Randomly generate the G-th generation variable speed limit population Chromosome s The ramp regulation rate real number sub-gene string ,in, express The corresponding s-th ramp regulation rate sub-gene in , and , thus obtaining the G-generation genetic population ,in, represents the sth individual, ; Step 6.3, get the Generation genetic population The sth individual The corresponding minimum fitness value Ramp regulatory subgene , so that the sth chromosome and Form the sth new individual, thus reaching the Generation genetic population The chromosome of each individual and its minimum fitness value are composed of the ramp regulation rate sub-genes New genetic population ; Step 6.3.1, the G generation genetic population The sth individual Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the traffic flow state quantity predicted by each cell of the Gth generation in the t+2th time period; Step 6.3.2: Substitute the traffic flow state predicted by each cell in the t+2 time period into the lower planning model, and use the inverse of the lower planning model as the fitness function to obtain the Generation genetic population The sth individual The fitness set ,in, express The corresponding Ramp-regulating subgenes The fitness value of Step 6.3.3, for Generation genetic population Middle Individual The fitness value set Compare the fitness values ​​of each ramp regulation rate sub-gene in the and select the minimum fitness value The corresponding ramp regulation rate sub-gene is recorded as ; Step 6.4, Generation genetic population The chromosomes of each individual and the ramp regulation rate sub-genes of each individual are crossover and mutation operations are performed to generate the first The offspring population of the generation population is , for the offspring population Perform the selection operation and get the Genetic population of the generation ; Step 6.5: Calculate the Genetic population of the generation The minimum fitness value corresponding to all individuals in ,in, express The minimum fitness value corresponding to the Mth individual in The sth chromosome in and The corresponding ramp regulation rate sub-gene Composition New genetic population The sth individual; like Less than , then retain the New genetic population The sth individual Otherwise, Assign to ,Will Assign to the sth new individual , thereby updating the New genetic population The sth new individual; Step 6.6, Assign to ,judge Is it true? If so, go to step 6.4; otherwise, it means it is completed The iterations get New genetic population , ; Step 6.7, Performing an annealing operation; Step 6.7.1, define the initial temperature as , termination temperature , temperature drop coefficient and the maximum number of cycles ; Define the current temperature as , and initialize ; The current number of cycles is , and initialize ;by As the current Generation population, denoted as ; Step 6.7.2, Input the improved cellular transmission model to conduct cooperative control simulation, and obtain the The traffic flow state quantity predicted by each cell is replaced by the inverse of formula (14) as the fitness function, thus obtaining The upper fitness value of each individual in the , and select the individual corresponding to the minimum fitness value as the first The best individual of the generation ,in, express Minimum fitness value The corresponding chromosomes express The corresponding ramp regulation rate sub-gene; Step 6.7.3, according to the Metropolis guidelines and ,right Each individual is selected and the first Generation population ; Step 6.7.4, judgment Is it true? If so, go to step 6.7.

5. Otherwise, Assign to , proceed to step 6.7.2; Step 6.7.5, if , then it means that the Optimal variable speed limit and optimal ramp adjustment rate under time period ,otherwise, Assign to , proceed to step 6.7.

3.

6. The dual-layer cooperative control method according to claim 5, characterized in that: The step 7 is to obtain the following steps when executing the ramp control strategy: Optimal ramp regulation rate for time period: Step 7.1, initialize the population size M, the current number of evolutions , maximum evolutionary generation , population crossover probability , population mutation probability ,initialization ; Step 7.2: Randomly generate M chromosomes in the t+1 time period, each chromosome represents a ramp adjustment rate, and obtain the The set of iterations is ,and , ; Step 7.3: Predict the first The ramp regulation rate corresponding to the sth chromosome in the time period is Traffic flow state quantity of each cell in the time period; Step 7.4, The sth chromosome in the subpopulation corresponds to the Substitute the traffic flow state quantity of the time period into the upper planning model, and take the inverse of the upper planning model formula (14) as the fitness function to obtain the first Fitness set of the generation population , Indicates the first Daidi The corresponding fitness value of each chromosome under the objective function; Step 7.5, for The set of fitness values ​​of all chromosomes Compare the fitness values ​​of each chromosome in Each generation of chromosomes with the minimum fitness value , indicating that this value is the strategy to minimize the traffic risk value of the merging area in the t+2 time period. The minimum fitness value is recorded as ; Step 7.6, the population size is M Generation population Perform crossover and mutation to generate Generation population The offspring population is , and its population size is also , for the offspring population Perform selection operations so that the population size is also The obtained chromosome is taken as The parent population of the generation , ; Step 7.7, calculate the The parent population of the generation The fitness value set of all chromosomes in , remember The minimum fitness value of each group of chromosomes in the generation is , and its corresponding chromosome is denoted as ;like Less than in step 7.5 , then the updated minimum fitness value is , and save its corresponding chromosome; otherwise, use The chromosome with the minimum fitness value in the generation chromosome Instead of The chromosome with the minimum fitness value among the first generation chromosomes enters the next generation; Step 7.8, Assign to ,judge Is it true? If so, go to step 7.6; otherwise, it means it is completed The final population set is recorded as , the population in The gene is the optimal solution, that is, the optimal ramp adjustment rate .

7. The dual-layer cooperative control method according to claim 5, characterized in that: The improved cellular transmission model predicts the traffic flow state of any generation in the t+2 time period according to the following steps: Step a: record the chromosome of any individual in any generation of population as ,definition ,in, represents the first variable speed limit control cell, representing the The variable speed limit value of cells, the ramp regulation rate sub-gene of any individual in any generation of population is recorded as ; Step b: Predict The sending and receiving capabilities of each cell in a time period; Step b1: predict the Time period The sending capacity of a variable rate limit control implementation cell and receiving capabilities ; (19) (20) In formula (19) to formula (20), , Respectively represent The variable speed limit control cell is The speed and the corresponding average density of the time period are obtained by formula (21); 、 Respectively represent The variable speed limit control cell is The average speed and average density of the time period are calculated using equations (7) to (10); (21) Calculate the sending and receiving capabilities of other mainline cells according to formula (1) to formula (2); Step b2: Predict the The sending capacity of the ramp control implementation cell under the time period ; (22) In formula (22), Indicates the Ramp adjustment rate corresponding to the time period; The receiving capacity of the ramp control implementation cell is calculated according to formula (4): ; Calculate the sending and receiving capabilities of other ramp cells according to formulas (3) and (4); Step c: Use formula (23) to determine Transmission flow of the mainline cell in time period i ; (23) When i=a, if , then using formula (24) we can get Otherwise, let ; (24) In formula (24), Indicates that the ramp control implementation cell is The sending capacity of the time period, Indicates the The confluence ratio between the mainline area and the ramp area in each time period, and ; Calculate the first order of other main line and ramp cells according to formula (5) to formula (6) Transmission flow rate in time period; Step d: Determine Traffic flow of ramp control implementation cells under time-varying speed limits : (25) Step e: Predict the first The average density of the main line cells in the time period i and the average density of the j-th ramp cell ; (26) (27) Step f: Predict the first The average speed of the main line cell in the time period i and the average speed of j ramp cells ; (28) (29)。 8. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the two-layer cooperative control method according to any one of claims 1 to 7, and the processor is configured to execute the program stored in the memory.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dual-layer cooperative control method according to any one of claims 1 to 7 are executed.

Images