Motorcade cluster formation control method, system and device based on multi-agent system

Abstract

A motorcade cluster formation control method, system and device based on a multi-agent system, and the method comprises the steps: initializing an existing motorcade cluster, obtaining a leading vehicle and a following vehicle in the motorcade cluster, endowing different role positioning vehicles with controllers, and forming a fixed motorcade cluster; preprocessing the motorcade cluster topology, modeling a lane change optimization problem by using an integer linear programming model, determining a specific target of the integer linear programming model, and determining constraint conditions of the integer linear programming model, so that the problem is solvable and conforms to a practical situation, and obtaining an optimized motorcade cluster topology; and according to the optimized fleet cluster topology, establishing a vehicle relation model, and according to a lane distribution result, sorting the vehicles according to the relative positions of the vehicles to complete the fleet cluster formation. The vehicle throughput of the bottleneck road section and safe and efficient lane changing of the vehicle are increased, and the lane changing cost of the vehicle is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of formation control of interconnected autonomous driving vehicles, and in particular to a vehicle fleet cluster formation control method, system and device based on a multi-agent system. Background Art

[0002] In real life, due to special scenarios such as road design, normal lane construction and maintenance, sudden car failures, and traffic accidents, lane reductions will occur on highways, which is a common bottleneck section. When there is such a single lane or even multiple lane reduction scenario on the highway, the highway's traffic capacity will be significantly reduced, and traffic congestion is likely to occur, leading to a series of problems such as increased traffic pollution, reduced safety levels, and longer travel time. The research subject of this paper is connected autonomous vehicles (CAVs). The formation composed of CAVs is regarded as a multi-agent system, and each CAV is regarded as an agent to study the formation consistency of the multi-agent system. Based on the multi-agent CAV formation method, model research and method design are carried out. By designing a vehicle dynamics controller that can meet the consistency protocol and through the ability of the agents to cooperate, control goals and tasks that are difficult for a single agent to complete are achieved.

[0003] Although the traditional vehicle following strategy IDM and the vehicle lane changing strategy SL2015 can drive normally by weighing the lane status of adjacent lanes on the road section and the vehicle's own status information, this method has obvious shortcomings. First, the judgment made by the vehicle from its own perspective is not the optimal solution in the vehicle group. The vehicle's lane changing decision may cause inconvenience to other vehicles, especially in an environment with reduced lanes. The decision made by the vehicle based on the traditional strategy may cause congestion of vehicles in a certain lane; secondly, the vehicle's adaptive lane changing causes the load conditions of different lanes to change. Changes in lane load conditions may cause other vehicles to change lanes unnecessary, increasing the vehicle's lane changing cost.

[0004] Therefore, a fleet cluster formation control method that can improve the vehicle throughput on bottleneck sections and reduce the lane changing cost of vehicles needs to be developed urgently. Summary of the invention

[0005] The object of the present invention is to provide a vehicle fleet cluster formation control method, system and device based on a multi-agent system, so as to solve at least one technical problem in the prior art.

[0006] The technical solution of the present invention is:

[0007] A fleet cluster formation control method based on a multi-agent system, comprising:

[0008] Initialize the existing fleet cluster, obtain the leading vehicle and the following vehicle in the fleet cluster, and assign controllers to vehicles with different role positioning to form a fixed fleet cluster;

[0009] Preprocess the fleet cluster topology, use the integer linear programming model to model the lane change optimization problem, determine the specific goal of the integer linear programming model, and determine the constraints of the integer linear programming model so that the problem is solvable and in line with the actual situation, and obtain the optimized fleet cluster topology;

[0010] According to the optimized fleet cluster topology, a vehicle relationship model is established. According to the results of lane allocation, the vehicles are sorted according to their relative positions to complete the fleet cluster formation.

[0011] Initializing the existing fleet cluster and obtaining the leading vehicle and the following vehicles in the fleet cluster includes:

[0012] Three key nodes are set up along the vehicle's travel direction, namely, the node P1 where the vehicle first accesses the roadside unit, the node P2 where the free vehicles form a single convoy cluster, and the location P3 where the lane is reduced;

[0013] At the beginning, if there is no vehicle running between P1 and P2, the roadside unit will regard the first vehicle passing P1 as the leader of the convoy cluster and the leader of the convoy in the lane where the vehicle is located in the current cluster; the vehicle that enters later is the following vehicle in the convoy cluster. If the vehicle is the first vehicle in the convoy in the lane, it will be regarded as the leader of the convoy.

[0014] When the leading vehicle of the convoy cluster passes the P2 node, the roadside unit will finalize the current convoy cluster to form a fixed convoy cluster.

[0015] The assigning controllers to positioning vehicles of different roles includes:

[0016] Grant the following vehicle:

[0017] Longitudinal controller of the following vehicle:

[0018]

[0019] Lateral controller of the following vehicle:

[0020]

[0021] Among them, i and L represent the leader of the follower and target convoys respectively; k represents the kth convoy cluster, k = 1,...,M; Represent the position, velocity, and control input of follower i respectively; They represent the neighbor vehicles of follower i in the target fleet and the neighbor vehicles in the current fleet respectively; Respectively represent the longitudinal position and longitudinal speed of the leading vehicle; α, β, γ, σ are non-negative constants; η m ∈{0, 1}, m∈{1, 2, 3}, selected according to the strategy; q i (t), w i (t) and r i They represent the lateral position, lateral velocity and expected distance between the lane-changing vehicle i and the leading vehicle L respectively; γ, σ are non-negative constants.

[0022] The assigning controllers to positioning vehicles of different roles includes:

[0023] The leadership vehicle is given:

[0024] Leading vehicle longitudinal controller:

[0025]

[0026] Leading vehicle lateral controller:

[0027]

[0028] Among them, Land represents a set of lanes; represents the set of vehicles in front and behind the lane where vehicle Lj is currently located; Lj represents the leading vehicle of the convoy on the jth lane; Ll represents the leading vehicle of the convoy on the lth lane; represents the longitudinal position of vehicle m, which is the last relative position in the k-1th convoy cluster at time t; represents the longitudinal expected distance between vehicle Lj and vehicle m; represents the longitudinal velocity of vehicle Lj, ​​m at time t; They represent the lateral position of the leading vehicle L1 of the convoy on lane l respectively; represents the lateral desired distance between vehicles Lj and Ll; represents the lateral velocity of vehicle Lj; represents the position of the target lane center line of vehicle Lj; η, η2, η3, η4, are all non-negative constants.

[0029] The specific objectives of the integer linear programming model include:

[0030] The average distribution of vehicles minimizes the difference in the number of vehicles in each lane; minimizes the lane change cost of the convoy cluster; and minimizes the number of lane changes of vehicles.

[0031] The constraint conditions of the integer linear programming model are determined, including:

[0032] Each vehicle can only be assigned to one lane:

[0033]

[0034] Lane Load Constraints:

[0035]

[0036] Minimize load differences in each lane:

[0037]

[0038] Lane change cost calculation:

[0039]

[0040] Among them, c ij is the number of lane changes of vehicle i from the original lane to lane j; T j is the number of vehicles on the jth lane; is the average number of vehicles; lane i Indicates the lane where vehicle i is currently located; x ij is a binary decision scalar indicating whether vehicle i is assigned to lane j.

[0041] The integer linear programming model comprises:

[0042]

[0043] Among them, c ij is the number of lane changes of vehicle i from the original lane to lane j; T j is the number of vehicles on the jth lane; is the average number of vehicles; x ij is a binary decision scalar, indicating whether vehicle i is assigned to lane j; ω ij is the number of lane changes of vehicle i weighted by priority; λ1, λ2, and λ3 are all weight coefficients.

[0044] A control system for a fleet cluster formation control method based on the multi-agent system as described above comprises:

[0045] An initialization module is used to initialize the existing fleet cluster, obtain the leading vehicle and the following vehicle in the fleet cluster, and assign controllers to vehicles with different role positioning to form a fixed fleet cluster;

[0046] The preprocessing module interacts with the initialization module for data preprocessing of the fleet cluster topology, models the lane change optimization problem using an integer linear programming model, determines the specific objectives of the integer linear programming model, and determines the constraints of the integer linear programming model, so that the problem is solvable and conforms to the actual situation, and obtains the optimized fleet cluster topology; establishes a vehicle relationship model, sorts the vehicles according to their relative positions according to the results of lane allocation, and completes the fleet cluster formation.

[0047] An electronic device for a fleet cluster formation, comprising:

[0048] Storage medium for storing computer programs;

[0049] The processing unit exchanges data with the storage medium, and is used to execute the computer program through the processing unit when performing a fleet cluster formation, and perform the steps of the fleet cluster formation control method based on the multi-agent system as described above.

[0050] A readable storage medium, comprising:

[0051] The readable storage medium stores a computer program; when the computer program is run, the steps of the vehicle fleet cluster formation control method based on the multi-agent system are executed as described above.

[0052] The beneficial effects of the present invention include at least:

[0053] The control method described in the present invention adopts a staged formation control strategy to generate a fleet cluster as a formation control unit, and a fixed number of vehicles facilitates the collective control of vehicles; at the same time, the topology of the fleet cluster is pre-processed and optimized using an integer linear programming model, and the resulting lane allocation can simultaneously meet the purpose of load balancing of continuous lanes and low-cost lane changing of the entire vehicle, thereby increasing the vehicle throughput of bottleneck sections and safe and efficient lane changing of vehicles; and, a more reasonable vehicle dynamics controller is provided. During the driving process, the vehicle must consider the specific information of the leading vehicle and the vehicles before and after the target fleet, and also the information of the vehicles before and after the current lane, so as to achieve safe lane changing in a multi-vehicle fleet scenario. At the same time, the designed controller meets the convergence and can form a stable fleet. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Comparison diagram of fleet cluster topology before and after preprocessing.

[0055] Figure 2 To simulate the motion trajectory of the vehicle.

[0056] Figure 3 is the velocity diagram of the vehicle along the x-axis.

[0057] Figure 4 is the velocity diagram of the vehicle along the y-axis.

[0058] Figure 5 This is a strategy diagram proposed by the present invention.

[0059] Figure 6 This is the traditional IDM+SL2015 strategy diagram. DETAILED DESCRIPTION

[0060] The present application is further described below in conjunction with the accompanying drawings.

[0061] The present invention is described in detail below with reference to the accompanying drawings and embodiments, but the practical application of the present invention is not limited to the following embodiments.

[0062] The purpose of the present invention is to overcome the problem of high traffic pressure on bottleneck sections in the case of reduced lanes in traditional vehicle control strategies, and to provide a platoon cluster formation control method for a multi-agent system. The method is divided into two stages, namely, the initialization stage of the platoon cluster and the coordinated lane-changing stage of the platoon cluster. In order to facilitate the formation control, all vehicles within a certain distance along the highway are regarded as a whole, and the vehicles in each lane are regarded as a platoon, thereby constructing a platoon cluster model (VPC) adapted to multi-lane scenarios; in subsequent strategies, the platoon cluster will be controlled as a unit. The vehicle information in the platoon cluster is collected, and the topology of the platoon cluster is preprocessed and optimized to achieve the purpose of lane load balancing and reducing the cost of vehicle lane change; then, a vehicle relationship model is established based on the preprocessed topology results, and a vehicle dynamics controller adapted thereto is designed to achieve safe driving of the vehicle; the vehicle throughput of the bottleneck section is improved and the lane-changing cost of the vehicle is reduced.

[0063] Based on this, the present invention provides the following embodiments. Specific embodiment 1:

[0065] The present invention provides an embodiment:

[0066] The convoy cluster formation control method of the multi-agent system proposed in this embodiment is divided into: the convoy cluster initialization stage and the convoy cluster coordinated lane changing stage. The convoy cluster initialization stage is mainly to organize the free vehicles into a formation control unit to facilitate the implementation of the subsequent lane changing strategy. The specific steps are as follows:

[0067] Step 1: Set up three key nodes along the vehicle's travel direction, namely, the node P1 where the vehicle first accesses the roadside unit, the node P2 where the free vehicles form a single fleet cluster, and the location P3 where the lane is reduced;

[0068] Step 2: Initially, if there is no vehicle between P1 and P2, the roadside unit will regard the first vehicle that passes P1 as the leader of the convoy cluster and the leader of the convoy in the lane where the vehicle is located in the current cluster; the vehicle that enters later is the following vehicle in the convoy cluster. If the vehicle is the first vehicle in the convoy in the lane, it will be regarded as the leader of the convoy;

[0069] Step 3: When the leading vehicle of the convoy cluster passes the P2 node, the roadside unit will finalize the current convoy cluster to form a fixed convoy cluster for the next stage of convoy cluster lane change control.

[0070] like Figure 1 Before the convoy cluster preprocessing, the first vehicle in each lane convoy is the leading vehicle, and the following vehicles are its following vehicles. Here, it is assumed that lanes 1 and 2 are continuous lanes, and lanes 3 and 4 are terminal lanes. After topological preprocessing, the lane allocation result is that L1, F1, F2, and F3 are allocated to lane 1, and L2, F4, F5, and F6 are allocated to lane 2. At the same time, the vehicles L1 and L2, which are ranked first in relative position, are defined as the leading vehicles of the two lane convoys.

[0071] The coordinated lane-changing stage of the convoy cluster is mainly to analyze and process the vehicle information of the convoy cluster to achieve the purpose of appropriate coordinated lane-changing; the specific steps are as follows:

[0072] Step 1: Preprocess the convoy cluster topology; use the integer linear programming model to model the lane change optimization problem, and propose three specific objectives of the model: (1) average allocation of vehicles so that the difference in the number of vehicles in each lane is minimized; (2) minimize the lane change cost of the convoy cluster; (3) minimize the number of lane changes of vehicles;

[0073] Step 2: Determine the constraints of the integer linear programming model so that the problem is solvable and consistent with the actual situation; add four constraints to the model: (1) Each vehicle can only be assigned to one vehicle; (2) Vehicle load constraint, the lane load refers to the number of vehicles assigned to the lane; (3) Minimize the load difference of each lane to ensure even distribution of vehicles; (4) The lane change cost is recorded as the number of lane changes of the vehicle, such as the lane change cost from lane 1 to lane 3 is 2;

[0074] Step 3: Solve the integer linear programming model and return the lane allocation plan for the vehicle; complete the preprocessing of the fleet cluster topology;

[0075] Step 4: Based on the optimized convoy cluster topology, a vehicle relationship model is established; according to the lane allocation results, the vehicles are sorted according to their relative positions. The vehicles at the front of the convoy in the topology are considered the leading vehicles of the convoy, and the vehicles at the back according to the topological structure are the following vehicles of the convoy.

[0077] The control strategy of the vehicle in the present invention adopts the design concept of leader-follower, and designs the lateral controller and longitudinal controller of the vehicle with different role positioning; specifically, it can be expressed as:

[0078] Longitudinal controller of the following vehicle:

[0079]

[0080] Lateral controller of the following vehicle:

[0081]

[0082] Among them, i and L represent the leader of follower i and target convoy respectively, k represents the kth convoy cluster, k=1,...,M. Represent the position, velocity, and control input of follower i respectively; They represent the neighbor vehicles of follower i in the target fleet and the neighbor vehicles in the current fleet respectively. Represent the longitudinal position and longitudinal velocity of the leading vehicle respectively. α, β, γ, σ are non-negative constants, η m ∈{0,1},m∈{1,2,3},,selected according to the strategy;q i (t), w i (t) and r i They represent the lateral position, lateral velocity and expected distance between the lane-changing vehicle i and the leading vehicle L respectively; γ, σ are non-negative constants.

[0083] Leading vehicle longitudinal controller:

[0084]

[0085] Leading vehicle lateral controller:

[0086]

[0087] Among them, Land represents the lane set, represents the set of vehicles in front and behind the lane where vehicle Lj is currently located; Lj represents the leading vehicle of the convoy on the jth lane, Ll represents the leading vehicle of the convoy on the lth lane, represents the longitudinal position of vehicle m, which is the last relative to the k-1th convoy cluster at time t, represents the longitudinal expected distance between vehicle Lj and vehicle m, represents the longitudinal velocity of vehicle Lj, ​​m at time t; They represent the lateral position of the leading vehicle L1 of the convoy on lane l, represents the lateral desired distance between vehicles Lj and Ll, represents the lateral velocity of vehicle Lj, represents the position of the target lane center line of vehicle Lj; η, η2, η3, η4, are all non-negative constants.

[0088] The vehicle formation control strategy is divided into the initialization phase of the convoy cluster and the coordinated lane-changing phase of the convoy cluster. The specific input and output of the convoy cluster initialization method are as follows:

[0089]

[0090]

[0091] The fleet cluster obtained by the above initialization method is used for vehicle platoon control in the next stage.

[0092] After generating a fixed fleet cluster, the topology of the fleet cluster needs to be preprocessed. This embodiment uses the integer linear programming (ILP) model to obtain the optimization method of vehicle allocation, and considers the advantages and disadvantages of the adjustment scheme from three optimization objectives, namely

[0093] 1) Vehicle average distribution: Minimize the difference in the number of vehicles in each lane;

[0094] 2) The total number of lane changes in the convoy cluster is the smallest: the lane change cost is the number of lane changes per vehicle c ij , so the goal is to minimize the total number of lane changes of all vehicles;

[0095] 3) Minimize the maximum number of lane changes of the vehicle;

[0096] In order to make the problem solvable and realistic, constraints need to be added to ensure that the number of cars per vehicle and per vehicle is reasonable. The constraints of the model are as follows:

[0097] 1) Each vehicle can only be assigned to one lane:

[0098]

[0099] 2) Lane load constraints:

[0100]

[0101] 3) Minimize the load difference of each lane:

[0102]

[0103] 4) Lane change cost calculation:

[0104]

[0105] The resulting integer linear programming model is as follows:

[0106]

[0107] The specific meanings of the parameters are as follows:

[0108] T j : The number of vehicles on the jth lane; The average number of vehicles; c ij : The number of lane changes of vehicle i from the original lane to lane j; x ij : Binary decision scalar, indicating whether vehicle i is assigned to lane j; ω ij : The number of lane changes of vehicle i according to the weighted priority; λ1, λ2, λ3 are weight coefficients used to balance the priorities between different objectives.

[0109] By solving the integer linear programming model, the target lane and optimized topology of each vehicle in the cluster can be obtained; by establishing a relationship model between vehicles based on the optimized topology, the leading vehicle and the following vehicle are allocated according to the target lane and the relative position of the vehicles.

[0110] like Figure 2 ,The vehicle motion trajectory diagram of the simulation experiment shows that the vehicles can be distributed to lanes 1 and 2 in an orderly and even manner, and form a stable fleet. Figure 3 , Figure 4 Respectively represent the longitudinal speed and lateral speed of each vehicle. The vehicle can maintain a stable speed after changing lanes, indicating that the vehicle's dynamic controller has convergence.

[0111] The steps to build a vehicle relationship model are as follows:

[0112]

[0113] After the vehicle relationship model is established, vehicles with different positioning will eventually reach the target lane under the guidance of the corresponding dynamic controller and complete the lane change operation.

[0114] Figure 5 , Figure 6 This is a partial screenshot of the simulation experiment comparison under the same vehicle data set. Figure 5 It can be seen that the vehicles in front form a convoy cluster and the vehicles can achieve an even distribution of lanes. Figure 6 In the traditional method, the lane load of the third lane is significantly higher than that of other lanes during the experiment. This is because vehicles in lanes 1 and 2 change to lane 3 before the lane reduction, while there are fewer vehicles changing to lane 4, resulting in congestion in lane 3.

[0115] The method described in this embodiment designs a two-stage vehicle formation control strategy for a lane reduction environment, which is divided into an initialization stage of the fleet cluster and a coordinated lane changing stage of the fleet cluster; and an integer linear programming model is designed for the fleet cluster topology preprocessing, which meets the three goals of lane load balancing, minimizing the overall lane changing cost of the fleet cluster, and minimizing the maximum number of lane changes of vehicles, and sets four model constraints that conform to the actual situation; finally, a lateral and longitudinal dynamics controller of the vehicle is designed for a multi-lane environment, which can achieve stable driving and coordinated lane changing of the vehicle. Specific embodiment 2:

[0117] The present invention also provides an embodiment:

[0118] A control system for a fleet cluster formation control method based on a multi-agent system as described in specific embodiment 1, comprising: an initialization module 100 and a preprocessing module 200; wherein the initialization module 100 is used to initialize the existing fleet cluster, obtain the leading vehicle and the following vehicle in the fleet cluster, and assign controllers to vehicles with different role positioning to form a fixed fleet cluster; the preprocessing module 200 interacts with the initialization module 100 for data preprocessing the fleet cluster topology, using an integer linear programming model to model the lane change optimization problem, determining the specific goal of the integer linear programming model, and determining the constraints of the integer linear programming model so that the problem is solvable and conforms to the actual situation, and an optimized fleet cluster topology is obtained; a vehicle relationship model is established, and vehicles are sorted according to their relative positions according to the results of lane allocation to complete the fleet cluster formation. Specific embodiment 3:

[0120] An electronic device for vehicle cluster formation, comprising: a storage medium and a processing unit; wherein the storage medium is used to execute the computer program through the processing unit when performing vehicle cluster formation, and perform the steps of the vehicle cluster formation control method of the multi-agent system as described in specific embodiment 1. Specific embodiment 4:

[0122] A computer-readable storage medium having a computer program stored therein; when the computer program is run, the computer program executes the steps of the vehicle fleet cluster formation control method of a multi-agent system as described in specific embodiment 1.

[0123] In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device, or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in combination with an instruction execution system, device, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0124] The above disclosures are only several specific implementation scenarios of the present invention, but the present invention is not limited thereto, and any changes that can be thought of by a person skilled in the art should fall within the protection scope of the present invention. The above serial numbers of the present invention are only for description and do not represent the advantages and disadvantages of the implementation scenarios.

Claims

1. A fleet cluster formation control method based on a multi-agent system, characterized in that: include: Initialize the existing fleet cluster, obtain the leading vehicle and the following vehicle in the fleet cluster, and assign controllers to vehicles with different role positioning to form a fixed fleet cluster; Preprocess the fleet cluster topology, use the integer linear programming model to model the lane change optimization problem, determine the specific goal of the integer linear programming model, and determine the constraints of the integer linear programming model so that the problem is solvable and in line with the actual situation, and obtain the optimized fleet cluster topology; According to the optimized fleet cluster topology, a vehicle relationship model is established. According to the results of lane allocation, the vehicles are sorted according to their relative positions to complete the fleet cluster formation.

2. The vehicle fleet formation control method based on a multi-agent system according to claim 1, characterized in that: Initializing the existing fleet cluster and obtaining the leading vehicle and the following vehicles in the fleet cluster includes: Three key nodes are set up along the vehicle's travel direction, namely, the node P1 where the vehicle first accesses the roadside unit, the node P2 where the free vehicles form a single convoy cluster, and the location P3 where the lane is reduced; At the beginning, if there is no vehicle running between P1 and P2, the roadside unit will regard the first vehicle passing P1 as the leader of the convoy cluster and the leader of the convoy in the lane where the vehicle is located in the current cluster; the vehicle that enters later is the following vehicle in the convoy cluster. If the vehicle is the first vehicle in the convoy in the lane, it will be regarded as the leader of the convoy. When the leading vehicle of the convoy cluster passes the P2 node, the roadside unit will finalize the current convoy cluster to form a fixed convoy cluster.

3. The vehicle fleet formation control method based on a multi-agent system according to claim 1, characterized in that: The assigning controllers to positioning vehicles of different roles includes: Grant the following vehicle: Longitudinal controller of the following vehicle: Lateral controller of the following vehicle: Among them, i and L represent the leader of the follower and target convoys respectively; k represents the kth convoy cluster, k = 1,...,M; Represent the position, velocity, and control input of follower i respectively; They represent the neighbor vehicles of follower i in the target fleet and the neighbor vehicles in the current fleet respectively; Respectively represent the longitudinal position and longitudinal speed of the leading vehicle; α, β, γ, σ are non-negative constants; η m ∈{0,1}, m∈{1,2,3}, selected according to the strategy; q i (t), w i (t) and r i They represent the lateral position, lateral velocity and expected distance between the lane-changing vehicle i and the leading vehicle L respectively; γ, σ are non-negative constants.

4. The vehicle fleet formation control method based on a multi-agent system according to claim 1, characterized in that: The assigning controllers to positioning vehicles of different roles includes: The leadership vehicle is given: Longitudinal controller of the leading vehicle: Leading vehicle lateral controller: Among them, Land represents a set of lanes; represents the set of vehicles in front and behind the lane where vehicle Lj is currently located; Lj represents the leading vehicle of the convoy on the jth lane; Ll represents the leading vehicle of the convoy on the lth lane; represents the longitudinal position of vehicle m, which is the last relative position in the k-1th convoy cluster at time t; represents the longitudinal expected distance between vehicle Lj and vehicle m; represents the longitudinal velocity of vehicle Lj, ​​m at time t; They represent the lateral position of the leading vehicle L1 of the convoy on lane l respectively; represents the lateral desired distance between vehicles Lj and Ll; represents the lateral velocity of vehicle Lj; represents the position of the target lane center line of vehicle Lj; η, η2, η3, η4, are all non-negative constants.

5. The vehicle group formation control method based on a multi-agent system according to claim 1 is characterized in that: The specific objectives of the integer linear programming model include: The average allocation of vehicles minimizes the difference in the number of vehicles in each lane; minimizes the lane change cost of the convoy cluster; and minimizes the number of lane changes of vehicles.

6. The vehicle fleet formation control method based on a multi-agent system according to claim 1, characterized in that: The constraint conditions of the integer linear programming model are determined, including: Each vehicle can only be assigned to one lane: Lane Load Constraints: Minimize load differences in each lane: Lane change cost calculation: Among them, c ij is the number of lane changes of vehicle i from the original lane to lane j; T j is the number of vehicles on the jth lane; is the average number of vehicles; lane i Indicates the lane where vehicle i is currently located; x ij is a binary decision scalar indicating whether vehicle i is assigned to lane j.

7. The vehicle fleet formation control method based on a multi-agent system according to claim 1, characterized in that: The integer linear programming model comprises: Among them, c ij is the number of lane changes of vehicle i from the original lane to lane j; T j is the number of vehicles on the jth lane; is the average number of vehicles; x ij is a binary decision scalar, indicating whether vehicle i is assigned to lane j; ω ij is the number of lane changes of vehicle i weighted by priority; λ1, λ2, and λ3 are all weight coefficients.

8. A control system based on the vehicle fleet formation control method of the multi-agent system according to any one of claims 1 to 7, characterized in that: include: An initialization module is used to initialize the existing fleet cluster, obtain the leading vehicle and the following vehicle in the fleet cluster, and assign controllers to vehicles with different role positioning to form a fixed fleet cluster; The preprocessing module interacts with the initialization module for data preprocessing of the fleet cluster topology, models the lane change optimization problem using an integer linear programming model, determines the specific objectives of the integer linear programming model, and determines the constraints of the integer linear programming model, so that the problem is solvable and conforms to the actual situation, and obtains the optimized fleet cluster topology; establishes a vehicle relationship model, sorts the vehicles according to their relative positions according to the results of lane allocation, and completes the fleet cluster formation.

9. An electronic device for a fleet of vehicles, comprising: Storage medium for storing computer programs; A processing unit exchanges data with the storage medium, and is used to execute the computer program through the processing unit when performing a fleet cluster formation, and perform the steps of the fleet cluster formation control method based on a multi-agent system as described in any one of claims 1 to 7.

10. A readable storage medium, characterized in that: include: The readable storage medium stores a computer program; when the computer program is running, it executes the steps of the fleet cluster formation control method based on a multi-agent system as described in any one of claims 1 to 7.

Images