A tree search-based merging sequence optimization method for mixed traffic flows on highways

Abstract

The present invention relates to the field of intelligent transportation, and in particular to a method for optimizing the merging sequence of mixed traffic flows on a highway based on tree search, comprising the following steps: S1, constructing a traffic scene of a two-lane main road and a single-lane ramp, and recording the current queue block merging sequence; S2, when a new queue block enters a control area, using tree search to obtain the optimal merging sequence containing the current queue block; S3, based on the optimal merging sequence obtained by the search, using the Pontryagin principle to solve the optimal driving trajectory of the vehicles in the queue block; S4, the head vehicle in the queue block drives according to the optimal driving trajectory, and the rest of the vehicles drive according to the following model to achieve merging. The present invention relies on the Internet of Vehicles technology to obtain the status information of the vehicle queue blocks in the main roads and ramps in the control area of ​​the highway in real time, and uses the controllability of CAV to complete the safe and efficient merging of the HDV in the queue block by searching for the optimal merging sequence and controlling the speed and lane change of the CAV.

Description

Technical Field

[0001] The invention relates to the field of intelligent transportation, and in particular to a method for optimizing a highway mixed traffic flow merging sequence based on tree search. Background Art

[0002] Thanks to its precise perception, communication, interaction and control capabilities, intelligent connected vehicles can effectively improve the accuracy and effectiveness of management and control strategies, inject new impetus into the development of intelligent transportation systems, and bring new ideas to solving the problem of highway ramp merging. Due to frequent lane changes and the randomness of driving behavior, the area near the entrance ramp often becomes the bottleneck area of ​​the highway, resulting in a decrease in the speed of vehicles in the highway network and an increase in driving time. Traditional human-driven vehicles trying to merge into the main road on the ramp cannot obtain real-time surrounding traffic conditions, resulting in stop-and-go phenomena and a higher risk of accidents.

[0003] Most existing methods for controlling traffic flow in merging areas of highways assume that all vehicles on the highway are intelligent connected vehicles, and only involve traffic scenarios where the main road is a single lane, ignoring the situation of multi-lane and mixed traffic flow, which is inconsistent with the actual scenario of highways. In addition, the perception uncertainty of traditional human driving leads to the need for driving guidance from intelligent connected vehicles instead of passively waiting for the merging gap. Summary of the invention

[0004] In view of the shortcomings of the existing single main road vehicle and ramp vehicle merging control method and the difficulty in using intelligent connected vehicles to optimally control mixed traffic flow, the present invention provides a highway mixed traffic flow merging sequence optimization method based on tree search.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for optimizing a highway mixed traffic flow merging sequence based on tree search comprises the following steps:

[0007] S1, constructs a traffic scenario of a two-lane trunk road and a single-lane ramp, divides the road into control areas and merging areas, organizes vehicles on the road into queue blocks, and the centralized controller is responsible for monitoring and controlling each queue block and recording the current queue block merging sequence;

[0008] S2, when a new queue block enters the control area, based on the previously recorded queue block merging sequence, with the driving efficiency, driving comfort, fuel consumption and pollution emissions of the vehicles in the current queue block as the target, a longitudinal driving cost function and a lateral lane change cost function are constructed, and a tree search is used to obtain the optimal merging sequence containing the current queue block;

[0009] S3, construct the vehicle kinematic model of all queue blocks, based on the searched optimal merging sequence, combined with the longitudinal cost function and the lateral cost function, and use the Pontryagin principle to solve the optimal driving trajectory of the vehicles in the queue block;

[0010] S4: The head car in the queue block drives according to the optimal driving trajectory, and the rest of the cars drive according to the following model to achieve merging.

[0011] Preferably, in S1, the queue block has an intelligent connected vehicle as the lead vehicle and the rest are driven by traditional people.

[0012] Preferably, in the queue block, the intelligent connected car as the head car is controlled by a centralized controller, and the rest of the people drive to follow the head car using the Newell following model.

[0013] Preferably, the two-lane trunk road includes a first trunk road L1 and a second trunk road L2, the first trunk road L1 is divided into a first control area section and a first merging area section, and a first merging point S1 is marked on the first merging area section; the second trunk road L2 is divided into a second control area section and a second merging area section, and a second merging point S2 is marked on the second merging area section; the single-vehicle ramp R1 is divided into a third control area section and a third merging area section;

[0014] When merging, vehicles passing through the first merging point S1 include vehicles traveling on the first main road L1 or vehicles changing lanes on the second main road L2; vehicles passing through the second merging point S2 include vehicles traveling on the first main road L1 or vehicles on the single-vehicle ramp R1.

[0015] Preferably, in S2, the process of obtaining the optimal merging sequence is: first, the action costs of all queue blocks after the new queue block is added are calculated, and then a depth-first search tree is used to find the optimal merging position including the current queue block and the optimal merging sequence of all queue blocks.

[0016] Preferably, in S3, during the construction of the vehicle kinematic models of all queue blocks, a safety following constraint is set between the head vehicles of two adjacent queue blocks on the same lane:

[0017]

[0018] in, Represents the minimum headway distance between two vehicles when following safely;

[0019] Represents a queue block The number of vehicles inside.

[0020] There are safe merging constraints between the head vehicles of two queue blocks on different lanes:

[0021]

[0022] in, Represents the length of the merge section;

[0023] Represents the expected speed of the queue block when it leaves the control area and enters the merging area.

[0024] Preferably, in S3, the optimal driving trajectory of the vehicle is determined by the optimal driving speed and optimal acceleration of the vehicle, and the acquisition method is as follows: first, a corresponding Hamiltonian function is constructed according to the longitudinal cost function and the lateral cost function, and then the Pontryagin maximum principle is used to solve the vehicle with driving efficiency, driving comfort, fuel consumption and pollution emissions as the goals.

[0025] A tree-search-based highway mixed traffic flow merging sequence optimization system comprises a data acquisition module, an upper tree search control module, a lower driving cost control module and a data output module. The data acquisition module is used to acquire the vehicle driving status of the remaining queue blocks in the control area and acquire the previously existing optimal queue merging sequence when a new queue block enters the control area on a two-car main road and a single-car ramp; the upper tree search control module is used to acquire the optimal merging sequence containing the current queue block by using tree search based on the previous queue block optimal merging sequence, considering the longitudinal driving cost and lateral lane-changing cost of the vehicles in the current queue block; the lower driving cost control module is used to construct a vehicle kinematic model of the queue block, and based on the optimal merging sequence, combined with the longitudinal cost function and the lateral cost function, the Pontryagin principle is used to solve the optimal driving trajectory of the vehicle; the data output module is used to output the optimal driving trajectory of the vehicle to the queue block, and the vehicle drives according to the optimal driving trajectory and merges.

[0026] A terminal device comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the steps of the method are implemented when the processor executes the computer program.

[0027] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method are implemented.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The present invention discloses a method for optimizing the merging sequence of mixed traffic flows on highways based on tree search. The method starts from the merging scene of multi-lane highways, is based on the idea that intelligent connected vehicles (CAVs) guide traditional human-driven vehicles (HDVs) to merge, relies on the vehicle networking technology to obtain the status information of the vehicle queue blocks in the main roads and ramps in the control area of ​​the highway in real time, utilizes the controllability of CAVs, searches for the optimal merging sequence through trees, and controls the speed and lane change of CAVs. Combined with the optimal control and vehicle kinematic model, the method takes the driving efficiency, fuel consumption, pollution emission and comfort of vehicles when merging as the goal, and uses the Pontryagin principle to solve the optimal driving trajectory of the vehicle to guide the HDV in the queue block to merge safely and efficiently. The method is applicable to different types of two-lane highways. Under the conditions of balanced and unbalanced traffic flow on highways, the method can effectively ensure the safety, efficiency, low fuel consumption and low pollution emission of vehicles when merging.

[0030] Furthermore, the expressway and ramp sections are divided into control areas and merging areas. Under the premise that the CAV is controllable, the HDVs and CAVs are organized into queues. The CAV serves as the head vehicle of each queue block, guiding the HDVs in the queue to merge safely and efficiently.

[0031] Furthermore, both the safe following constraint and the safe merging constraint are intended to ensure the safe merging of vehicle queue blocks.

[0032] Furthermore, the longitudinal driving cost function takes into account the driving efficiency, fuel consumption, comfort and pollution emissions of the queue block, and the lateral cost function takes into account the comfort of the vehicle when changing lanes to ensure a smooth and safe lane change process.

[0033] The present invention provides a highway mixed traffic flow merging sequence optimization system based on tree search, which provides a tree search algorithm at the upper layer to dynamically search for the optimal merging sequence of vehicle queues, and provides an optimal collaborative merging control method at the lower layer to determine the optimal driving trajectory of the vehicle queue based on the merging sequence calculated by the upper layer. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of a method for optimizing a merging sequence of mixed traffic flows on a highway based on tree search according to the present invention;

[0035] Figure 2 Schematic diagram of a highway traffic scene in an embodiment.

[0036] Figure 3 It is a schematic diagram of a hierarchical optimization framework in an embodiment of the present invention.

[0037] Figure 4 Schematic diagram of a tree search result in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described in detail below in conjunction with specific embodiments, which are intended to explain the present invention rather than to limit it.

[0039] The present invention discloses a method for optimizing the merging sequence of mixed traffic flows on expressways based on tree search. Figure 1 , including the following steps:

[0040] S1, constructs a traffic scenario of a two-lane trunk road and a single-lane ramp, divides the road into control areas and merging areas, organizes vehicles on the road into queue blocks, and the centralized controller is responsible for monitoring and controlling each queue block and recording the current queue block merging sequence. The details are as follows:

[0041] The main highway and ramp sections are segmented (e.g. Figure 2 As shown in the figure, the two-vehicle main road includes the first main road L1 and the second main road L2. The first main road L1 is divided into a first control area section and a first merging area section, and the first merging point S1 is marked on the first merging area section; the second main road L2 is divided into a second control area section and a second merging area section, and the second merging point S2 is marked on the second merging area section; the single-vehicle ramp R1 is divided into a third control area section and a third merging area section.

[0042] When vehicles enter the control area, they are organized into queue blocks, each of which contains HDVs guided by CAVs, that is, the queue block has an intelligent connected vehicle as the lead vehicle and the rest are driven by traditional people.

[0043] In addition, the centralized controller assigns unique IDs to the queue blocks and the CAVs and HDVs therein, corresponding to , , ,The centralized controller continuously sends control commands to the queue block head vehicle CAV in the control area through the roadside equipment unit (RSU) according to the queue block status, until it leaves the control area and enters the merging area.

[0044] During merging, vehicles passing through the first merging point S1 include vehicles traveling on the first main road L1 or vehicles changing lanes on the second main road L2; vehicles passing through the second merging point S2 include vehicles traveling on the first main road L1 or vehicles on the single-vehicle ramp R1. The main road queue blocks are respectively the first main road queue block D1 and the second main road queue block D2 according to the lanes, wherein the vehicles in the first main road queue block D1 keep driving in the current lane, and the vehicles in the second main road queue block D2 change lanes according to the longitudinal driving cost and the lateral lane change cost. The ramp queue block keeps driving in the current lane or merges to the second lane L2 according to the longitudinal driving cost.

[0045] S2, when a new queue block enters the control area, based on the previously recorded queue block merging sequence, with the driving efficiency, driving comfort, fuel consumption and pollution emissions of the vehicles in the current queue block as the target, the longitudinal driving cost function and the lateral lane change cost function are constructed, and the optimal merging sequence containing the current queue block is obtained by tree search. That is, when a new vehicle queue block comes in, the possible merging position of the current newly entered queue block is calculated based on the known previous vehicle queue merging sequence, and then the merging sequence is obtained according to the merging position, and then the new vehicle queue block is allowed to travel at a predetermined speed according to the solved merging sequence.

[0046] The process of obtaining the optimal merging sequence is as follows: First, calculate the action costs of all queue blocks after the new queue block is added, and then use the depth-first search tree to find the optimal merging position containing the current queue block and the optimal merging sequence of all queue blocks. In order to ensure the safe and efficient merging of queue blocks, it is necessary to pre-search the optimal merging sequence for each merging point. Since the queue blocks on lane L2 can choose to change lanes to lane L1 or continue driving on L2, the merging sequence of merging points S1 and S2 is only affected by the merging position of the queue blocks in lane L2. Figure 2 For example, refer to the scene in Figure 3 ,Consider queue block B4. Since B4 can be part of the merging sequence of merging points S1 or S2, different actions of B4, i.e., keeping the current lane or changing lanes to L1, will incur specific costs. Therefore, minimizing the global cost in the control area requires determining the expected merging position of B4, i.e., determining whether B4 needs to change lanes, and its position before / after changing lanes.

[0047] The optimal merging sequence for each merging point can be calculated using a centralized controller. The specific tree search algorithm steps are as follows:

[0048] Step (1) Input the known confluence sequence of the current confluence points S1 and S2, recorded as sets M and N, and proceed to step (2).

[0049] Step (2) Initialize the parameters, initialize the root node of the tree, set its value to MS, initialize the two child nodes S1 and S2 of the root node, corresponding to the two confluence points respectively, and continue to step (3).

[0050] Step (3) assigns the right child node of the S1 node to M, assigns the left child node of the S2 node to N, and continues with step (4).

[0051] Step (4) Whenever a new vehicle queue block enters the control area, the following steps are executed in a loop:

[0052] Step (4a) assigns a unique identifier to the vehicle queue block ,judge The lane to which it belongs, if From lane L2, go to step (4d); if Coming from lane L1, then proceed to step (4b); if Coming from lane R1, proceed to step (4c).

[0053] Step (4b) updates the value of set M by Add it to the set M, update the value of the right child node of the S1 node, and return to step (4) to continue execution.

[0054] Step (4c) updates the value of the set N, Add it to the set N, update the value of the left child node of the S2 node, and return to step (4) to continue execution.

[0055] Step (4d) For all vehicle queue blocks in set M, calculate the corresponding cumulative sum of longitudinal travel costs, denoted as ; For all vehicle queue blocks in set N, calculate the corresponding cumulative sum of longitudinal travel costs, recorded as ;calculate The corresponding lane-changing cost is ;if ,but You should switch from lane L2 to lane L1 and continue with step (4e); otherwise, if ,but You should continue driving in lane L2 and proceed to step (4f).

[0056] Step (4e) After the lane change, the potential merging positions and corresponding merging sequences on lane L1 are added to the descendant nodes and leaf nodes of node S1. The cost corresponding to the merging sequence in each leaf node is traversed and calculated, and the one with the lowest cost value is selected as the optimal merging sequence for the current merging point S1 and output. The optimal merging sequence for the current merging point S2 is output as set N, and the process returns to step (4) to continue.

[0057] Step (4f) The potential merging positions and corresponding merging sequences on lane L2 are added to the descendant nodes and leaf nodes of node S2. The cost corresponding to the merging sequence in each leaf node is traversed and calculated. The one with the lowest cost value is selected as the optimal merging sequence for the current merging point S2 and output. The optimal merging sequence for the current merging point S1 is output as set M, and the process returns to step (4) to continue.

[0058] like Figure 4 As shown, Figure 4 is based on Figure 1An example of a tree search result constructed for the scenario in . At this time, for the confluence point S1, the optimal confluence sequence is {B4->B6->B3}, and for the confluence point S2, the optimal confluence sequence is {B5->B2->B1}.

[0059] S3, constructs the vehicle kinematic model of all queue blocks, based on the optimal merging sequence obtained by search, combined with the longitudinal cost function and the lateral cost function, and uses the Pontryagin principle to solve the optimal driving trajectory of the vehicles in the queue block.

[0060] In the queue block, the intelligent connected car as the lead vehicle is controlled by a centralized controller, and the rest of the people drive using the Newell following model to follow the lead vehicle.

[0061] Considering that each vehicle can be regarded as a mass point moving along a designated lane, and only the head CAV in a vehicle platoon block can be controlled by a centralized controller, the head CAV in the platoon block The third-order motion state equation can be expressed as:

[0062]

[0063] in, represent Moment CAV location;

[0064] represent Moment CAV speed;

[0065] represent Moment CAV acceleration;

[0066] represent Moment CAV The acceleration of

[0067] Due to the speed limit on the highway and the vehicle's own dynamic performance and mechanical structure characteristics, the vehicle's motion state and control input are constrained by boundary conditions, and there is a formula:

[0068]

[0069] in, represents the minimum speed limit;

[0070] represents the maximum speed limit;

[0071] Represents the minimum acceleration limit;

[0072] Represents the maximum acceleration limit;

[0073] Represents the minimum jerk limit;

[0074] Represents the maximum jerk limit;

[0075] Representing CAV Time of entry into the controlled area;

[0076] Representing CAV The time to leave the control area and enter the merging area.

[0077] With the head car CAV in the queue block Since the HDV in the queue block follows the lead vehicle in the following car model, the Newell following car model can be used to estimate the queue block. Within HDV motion state, then HDV exist The position and velocity at time can be expressed as:

[0078]

[0079] in, represent Moment HDV location;

[0080] represent Moment HDV speed;

[0081] HDV Response time;

[0082] HDV Minimum following distance;

[0083] Represents a queue block The number of vehicles.

[0084] On the same lane, there is a safety following constraint between the leading vehicles of two adjacent queue blocks:

[0085]

[0086] in, Represents the minimum headway distance between two vehicles when following safely; Represents a queue block The number of vehicles; Representing CAV Time to leave the control area and enter the merging area

[0087] There are safe merging constraints between the head vehicles of two queue blocks on different lanes:

[0088]

[0089] in, Represents the length of the merge section;

[0090] Represents the expected speed of the queue block when it leaves the control area and enters the merging area.

[0091] In order to comprehensively evaluate the driving and lane-changing decisions of the vehicle queue, the longitudinal and lateral driving cost functions are considered to be decoupled, and the longitudinal driving cost function and the lateral lane-changing cost function are established. The longitudinal driving cost function takes into account the driving efficiency, fuel consumption, comfort and pollution emissions of the queue, which can be expressed as follows:

[0092]

[0093] in, Represents the weight coefficient of the queue block driving efficiency item;

[0094] Represents the weight coefficient of the queue block fuel consumption item;

[0095] Represents the weight coefficient of the driving comfort item of the queue block;

[0096] represents the weight coefficient of the pollution emission item of the queue block;

[0097] represents a constant value, which is 0.004;

[0098] Representative platoon block vehicle CAV Longitudinal travel speed;

[0099] Representative platoon block vehicle CAV Longitudinal driving acceleration;

[0100] Representative platoon block vehicle CAV Longitudinal driving acceleration.

[0101] In the lateral cost function, the comfort of the vehicle when changing lanes is taken into account to ensure the smoothness and safety of the lane change process, which is expressed as:

[0102]

[0103] in, and represents the weight coefficient;

[0104] Representative platoon block vehicle CAV lateral acceleration;

[0105] Representative platoon block vehicle CAV The lateral acceleration;

[0106] Represents a queue block The lane change start time;

[0107] Represents a queue block The lane change end time.

[0108] The Pontryagin maximum principle is used to solve the optimal driving speed and acceleration of the queue block, and the corresponding Hamiltonian function is constructed for the longitudinal and transverse cost functions. , there is a longitudinal Hamiltonian to solve the equation:

[0109]

[0110] in, represents the Lagrange multiplier; Represents the expected speed of the queue block when it leaves the control area and enters the merging area.

[0111] The corresponding Euler-Lagrange equation is:

[0112]

[0113] in, Representative platoon block vehicle CAV The vertical position of the

[0114] Solving the Euler-Lagrange equation yields:

[0115]

[0116] in,

[0117]

[0118] For parameters , can be calculated by solving the following matrix equation:

[0119]

[0120]

[0121] in, Representing CAV The longitudinal position when leaving the control area and entering the merging area;

[0122] Representing CAV Longitudinal speed when leaving the control area and entering the merging area;

[0123] Representing CAV Longitudinal acceleration when leaving the control area and entering the merging area.

[0124] Correspondingly, for the head CAV in the queue block , there is a transverse Hamiltonian to solve the equation:

[0125]

[0126] in, represents the Lagrange multiplier;

[0127] Representing CAV Lateral speed when changing lanes.

[0128] There are Euler-Lagrange equations:

[0129]

[0130] Solved:

[0131]

[0132] in, , , for the parameter , can be calculated by solving the following matrix equation:

[0133]

[0134] in, Representing CAV lateral position before executing a lane change;

[0135] Representing CAV lateral speed before executing a lane change;

[0136] Representing CAV lateral acceleration before executing a lane change;

[0137] Representing CAV lateral position after executing a lane change;

[0138] Representing CAV lateral speed after executing a lane change;

[0139] Representing CAV Lateral acceleration after executing a lane change.

[0140]

[0141] S4: The head car in the queue block drives according to the optimal driving trajectory, and the rest of the cars drive according to the following model to achieve merging.

[0142] In the present invention, when a new queue block enters the control area, it is added to the centralized controller, and the driving status of the remaining queue blocks in the controller and the previously existing optimal merging sequence of the queue blocks are obtained; then based on the existing optimal merging sequence of the queue block, the longitudinal driving cost function and the lateral lane change cost function are constructed with the driving efficiency, driving comfort, fuel consumption and pollution emission of the vehicles in the current queue block as the target, and the optimal merging sequence containing the current queue block is obtained by tree search. Then, the vehicle kinematic model of the current queue block is constructed, and the optimal driving trajectory of the vehicles in the queue block is solved by the Pontryagin principle based on the optimal merging sequence obtained by the search, combined with the longitudinal cost function and the lateral cost function; the vehicles in the current queue block drive according to the optimal driving trajectory to merge. Relying on the Internet of Vehicles technology, the status information of the vehicle queue blocks in the main roads and ramps in the control area of ​​the expressway is obtained in real time, and the controllability of CAV is utilized to search for the optimal merging sequence and control the speed and lane change of CAV to guide the HDV in the queue block to merge safely and efficiently.

[0143] The present invention also discloses a highway mixed traffic flow merging sequence optimization system based on tree search, comprising a data acquisition module, an upper tree search control module, a lower driving cost control module and a data output module, wherein the data acquisition module is used to obtain the vehicle driving status of the remaining queue blocks in the control area when a new queue block enters the control area on a two-car main road and a single-car ramp, and obtain the previously existing optimal queue merging sequence; the upper tree search control module is used to obtain the optimal merging sequence containing the current queue block by tree search based on the previous queue block optimal merging sequence, considering the longitudinal driving cost and lateral lane change cost of the vehicles in the current queue block; the lower driving cost control module is used to construct a vehicle kinematic model of the queue block, based on the optimal merging sequence, combined with the longitudinal cost function and the lateral cost function, and using the Pontryagin principle to solve the optimal driving trajectory of the vehicle; the data output module is used to output the optimal driving trajectory of the vehicle to the queue block, and the vehicle drives according to the optimal driving trajectory and merges.

[0144] The present invention also discloses a terminal device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the steps of the method are implemented when the processor executes the computer program.

[0145] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method are implemented.

[0146] The above description is only a preferred embodiment of the present invention and is not intended to impose any limitation on the technical solution of the present invention. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can also be subjected to several simple modifications and substitutions, and these modifications and substitutions are also within the scope of protection covered by the claims.

Claims

1. A method for optimizing the merging sequence of mixed traffic flows on highways based on tree search, characterized in that: The following steps are involved: S1, constructs a traffic scenario of a two-lane trunk road and a single-lane ramp, divides the road into control areas and merging areas, organizes vehicles on the road into queue blocks, and the centralized controller is responsible for monitoring and controlling each queue block and recording the current queue block merging sequence; S2, when a new queue block enters the control area, based on the previously recorded queue block merging sequence, with the driving efficiency, driving comfort, fuel consumption and pollution emissions of the vehicles in the current queue block as the target, a longitudinal driving cost function and a lateral lane change cost function are constructed, and a tree search is used to obtain the optimal merging sequence of all queue blocks including the current queue block; S3, construct the vehicle kinematic model of all queue blocks, based on the searched optimal merging sequence, combined with the longitudinal cost function and the lateral cost function, and use the Pontryagin principle to solve the optimal driving trajectory of the vehicles in the queue block; S4: The head car in the queue block drives according to the optimal driving trajectory, and the rest of the cars drive according to the following model to achieve merging.

2. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 1, characterized in that: In S1, the queue block is headed by a smart connected car and the rest are driven by traditional people.

3. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 2, characterized in that: In the queue block, the intelligent connected car as the lead vehicle is controlled by a centralized controller, and the rest of the people drive using the Newell following model to follow the lead vehicle.

4. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 2, characterized in that: The two-lane trunk road includes a first trunk road L1 and a second trunk road L2, the first trunk road L1 is divided into a first control area section and a first merging area section, and a first merging point S1 is marked on the first merging area section; The second trunk road L2 is divided into a second control area section and a second merging area section, and a second merging point S2 is marked on the second merging area section; the single-vehicle ramp R1 is divided into a third control area section and a third merging area section; During merging, vehicles passing through the first merging point S1 include vehicles traveling on the first trunk road L1 or vehicles changing lanes on the second trunk road L2; The vehicles passing through the second merging point S2 include vehicles traveling on the first trunk road L1 or vehicles on the single-vehicle ramp R1.

5. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 4 is characterized in that: In S2, the process of obtaining the optimal merging sequence is as follows: first, the action costs of all queue blocks after the new queue block is added are calculated, and then the optimal merging position of the current queue block and the optimal merging sequence of all queue blocks are found using a depth-first search tree.

6. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 5, characterized in that: In S3, when constructing the vehicle kinematic models of all queue blocks, a safety following constraint is set between the head vehicles of two adjacent queue blocks on the same lane: in, Represents the minimum headway distance between two vehicles when following safely; Represents a queue block The number of vehicles; Representing CAV The time of leaving the control area and entering the merging area; There are safe merging constraints between the head vehicles of two queue blocks on different lanes: in, Represents the length of the merging area section; Represents the expected speed of the queue block when it leaves the control area and enters the merging area.

7. The method for optimizing the merging sequence of mixed traffic flows on highways based on tree search according to claim 6, characterized in that: In S3, the optimal driving trajectory of the vehicle is determined by the optimal driving speed and optimal acceleration of the vehicle, which can be obtained as follows: first, the corresponding Hamiltonian function is constructed according to the longitudinal cost function and the lateral cost function, and then the Pontryagin maximum principle is used to solve the problem with the vehicle's driving efficiency, driving comfort, fuel consumption and pollution emissions as the goals.

8. A system based on the highway mixed traffic flow merging sequence optimization method based on tree search according to any one of claims 1 to 7, characterized in that: The system comprises a data acquisition module, an upper tree search control module, a lower driving cost control module and a data output module. The data acquisition module is used to obtain the vehicle driving status of the remaining queue blocks in the control area and the previously existing optimal queue merging sequence when a new queue block enters the control area on a two-car main road and a single-car ramp; the upper tree search control module is used to obtain the optimal merging sequence containing the current queue block by tree search based on the previous optimal merging sequence of the queue block, considering the longitudinal driving cost and lateral lane change cost of the vehicles in the current queue block; the lower driving cost control module is used to construct a vehicle kinematic model of the queue block, based on the optimal merging sequence, combined with the longitudinal cost function and the lateral cost function, and using the Pontryagin principle to solve the optimal driving trajectory of the vehicle; the data output module is used to output the optimal driving trajectory of the vehicle to the queue block, and the vehicle drives according to the optimal driving trajectory and merges.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images