Intelligent connected vehicle multi-lane formation method and system without fixed configuration

Abstract

The application provides a multi-lane formation method and system for intelligent networked vehicles without fixed configuration, which controls vehicles to enter target lanes to form a vehicle fleet according to expected lane selection information when the vehicles pass through a conflict area; a neighbor vehicle subset of the vehicles is constructed by analyzing the interaction influence relationship among the vehicles; a flexible and variable vehicle fleet configuration is realized according to the neighbor vehicle subset defined non-fixed following relationship and collision avoidance constraints; the multi-vehicle trajectory planning problem in the formation process is defined as a mixed integer linear programming model, the decision variables simultaneously consider the vehicle lateral and longitudinal motion, and the target function optimizes the lane deviation and driving speed changing with time; the formation is carried out by using a rolling horizon optimization and control framework, the vehicle state is updated at each sampling time, the multi-vehicle trajectory planning problem in the planning horizon is solved, and the decision result and control quantity are executed. The application effectively improves the flexibility, maneuverability and formation efficiency of the formation in the traffic flow, and improves the traffic flow passing speed of the conflict area.

Description

Technical Field

[0001] This invention relates to the field of collaborative decision-making and control technology for multi-vehicle cooperative driving, specifically to a method and system for multi-lane platooning of intelligent connected vehicles without fixed configuration, and also to a corresponding computer terminal and computer-readable storage medium. Background Technology

[0002] Traffic congestion is a persistent problem in transportation networks, particularly pronounced in conflict zones. Conflict zones are bottlenecks in traffic flow interaction, such as intersections, highway entrance and exit ramps, and work areas with reduced lane numbers. Frequent lane changes, overtaking, and yielding within conflict zones often cause traffic flow oscillations, inducing accidents and congestion. The development of intelligent connected vehicle technology offers a new approach to solving congestion in conflict zones. Under intelligent connectivity, vehicle platooning, through cooperative driving, unifies the orderly yielding, overtaking, and lane changes of vehicles within the same lane, forming queues on the target lanes. This can improve traffic efficiency in conflict zones, optimize traffic flow speed, smooth traffic dynamics, and prevent congestion.

[0003] Formation control methods widely used in the field of mobile robots or drones, such as virtual structure methods, are difficult to apply to structured road environments and complex traffic rules. Furthermore, structured topologies are difficult to adapt to dynamically changing and highly uncertain traffic environments, and rigid formation structures are not conducive to adaptability, thus limiting the flexibility and maneuverability of the formation process. Behavior-based or artificial potential field-based methods are easy to solve problems such as formation, obstacle avoidance, and reconfiguration of formations, but they are difficult to describe the overall behavior of the system and also difficult to consider the global performance indicators of the formation process.

[0004] Currently, the most widely used platooning method in the transportation sector is the linear platooning of the leader-follower approach. This type of platooning method is well-suited for road traffic environments with relatively stable and singular directions of travel, helping to improve traffic performance. However, most existing studies only focus on the formation and maintenance of platoons on single lanes, neglecting the lane-changing issues involved in platooning in multi-lane environments. Furthermore, the front-to-back positions and following relationships of vehicles within a platoon, i.e., the platoon configuration, are mostly predetermined. When forming a platoon, a trajectory is planned for each vehicle to enter its pre-determined position within the platoon, controlling the vehicles to reach their respective target positions to complete the platooning. However, this pre-fixed configuration is not conducive to coping with dynamically changing traffic environments, limiting the flexibility and maneuverability of the platooning process, and making it difficult to maximize the optimization effect of vehicle platooning on traffic flow. Summary of the Invention

[0005] To address the aforementioned shortcomings in the prior art, this invention provides a method and system for multi-lane platooning of intelligent connected vehicles without a fixed configuration, and also provides a corresponding computer terminal and computer-readable storage medium.

[0006] According to one aspect of the present invention, a method for multi-lane platooning of intelligent connected vehicles without fixed configuration is provided, comprising:

[0007] Obtain the target lane information of vehicles entering the platooning area, and collect the current status information of the vehicles and the overall traffic flow information;

[0008] Based on the target lane information, vehicle status information, and traffic flow status information, a linear dynamic model of the vehicle is established. This model is used to describe the vehicle's longitudinal following behavior and lateral lane change decision.

[0009] All vehicles are numbered according to the order in which they arrive at the platooning area. Based on the lanes that a vehicle needs to pass through to complete the platooning, a subset of neighboring vehicles affected by its following and lane-changing behaviors is constructed for each vehicle's linear dynamic model. For each vehicle and the vehicles in its neighboring vehicle subset, non-fixed following relationships and collision avoidance constraints are defined to construct an optimizable configuration for multi-lane vehicle platooning.

[0010] Based on the optimizable configuration of the multi-lane vehicle formation, the trajectory planning problem of the multi-lane vehicle formation is defined as a mixed integer linear programming model and solved to obtain the optimized trajectory planning result.

[0011] The longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning result are sent to the vehicle. The vehicle executes the corresponding lateral lane change decision and longitudinal control quantity, and the lateral lane change trajectory of the vehicle is calculated at the vehicle end.

[0012] The system re-collects the current state information of the vehicles and updates the state vector. It also updates the subset of neighboring vehicles, plans to move forward one step in the time domain and solves the trajectory planning problem for multi-lane vehicle formation. The system obtains the decision control quantity for the current moment, issues it to the vehicles, and executes it. This process is repeated until all vehicles enter the target lane and the formation is completed.

[0013] Preferably, the step of acquiring the target lane information of vehicles entering the platooning area and collecting the current status information of the vehicles and the overall traffic flow information includes:

[0014] Inspect all vehicles entering the platoon area;

[0015] Obtain the target location and path of each vehicle and match them with the travel direction of each lane to obtain the desired lane information for the vehicle to pass through the conflict zone, i.e., the target lane information:

[0016] Collect current status information of all vehicles, including vehicle position, speed, and acceleration;

[0017] Obtain overall traffic flow information, which includes current average traffic speed and vehicle density information.

[0018] Preferably, the step of establishing a linear dynamic model of the vehicle based on the target lane information, vehicle status information, and traffic flow status information, which is used to describe the vehicle's longitudinal following behavior and lateral lane-changing decisions, includes:

[0019] A vehicle linear dynamic model is established, wherein the length of the time window provided for a vehicle to complete a lane change is determined based on the traffic flow status information and used as a parameter of the vehicle linear dynamic model, and the current time state variable of the vehicle linear dynamic model is updated using the vehicle status information.

[0020] The longitudinal following behavior of the vehicle is established using an incremental second-order linear model:

[0021]

[0022] v i (k+1)=v i (k)+a i (k-1)Δt+Δa i (k)Δt,

[0023] a i (k)=a i (k-1)+Δa i (k),

[0024] Where, p i (*) and v i (*) represent the longitudinal position and velocity of vehicle i at the * discrete time, respectively, and a i (*) represents the acceleration maintained by vehicle i during the *th discrete period, Δt represents the length of the discrete time interval, and Δa i (*) indicates the change in acceleration at the *th discrete time.

[0025] The lateral lane change decision part is established using a time-delay model that describes the time taken for vehicles to change lanes and has undergone state augmentation processing:

[0026]

[0027]

[0028] Where, n i (*) indicates the decision state of vehicle i at time * regarding which lane it plans to go to. Let N represent the rightward and leftward lane change decisions made by vehicle i at time *. c This indicates the length of the time window provided for a vehicle to complete a lane change.

[0029] The state vector x of the vehicle linear dynamic model i (k) is:

[0030] x i (k)=[p i (k),v i (k),a i (k-1),n i (kN c ),n i (kN c +1),...,n i (k-1)] T

[0031] The decision input u of the vehicle linear dynamic model i (k) is:

[0032]

[0033] Preferably, determining the length of the time window provided for the vehicle to complete the lane change includes:

[0034] Based on the real-time traffic flow status information, calculate the sufficient time required for vehicles to perform lane-changing maneuvers:

[0035] t c =exp(c v v flow +c d d traffic +c0)

[0036] Among them, v flow d represents the average speed of the traffic flow within the current formation area. traffic Vehicle density within the formation area; c v c d c0 and c0 are parameters used to calculate lane change time, respectively. In the offline phase, multiple linear regression is used based on experimental and collected road vehicle lane change data. We obtain, where X n Let ψ be a vector composed of the average speed of traffic flow and the vehicle density in the observation n, and let ψ be the corresponding parameter, ∈ n For the error term of the observed value n;

[0037] Online phase calculation t c And according to The length of the time window provided for the vehicle to complete the lane change is obtained.

[0038] Preferably, constructing the subset of neighboring vehicles includes:

[0039] Construct the set of lanes that all vehicles i∈I must take to reach the target lane from their current position. or Where 'l' represents the lane number. Let i be the target lane number; for each vehicle i, obtain the subset of vehicles whose lanes intersect with its lanes on the lateral lane change. in, The elements are derived from the set I of all vehicles participating in the formation. For vehicles The set of lanes that must be taken to reach its target lane; based on the current longitudinal distance between vehicles, a subset of neighboring vehicles that may be affected by the following and lane-changing behavior of vehicle i in the prediction time domain. in, For vehicles The longitudinal position, d LIS It is the distance parameter affected by vehicle following and lane changing behavior.

[0040] Preferably, the definition of the non-fixed following relationship and collision avoidance constraint may further include the following operations:

[0041] For vehicle i and all vehicles in the subset of vehicle i's neighboring vehicles. Define three types of auxiliary binary variables:

[0042]

[0043]

[0044]

[0045] in, and Indicate vehicle i and vehicle k at time k respectively The front-back and left-right positional relationships; Indicate vehicle i and vehicle Is the lateral distance less than one lane? This indicates the vehicle's position in the transverse lane. The lane change progress coefficient is obtained from the lateral trajectory planner of the vehicle layer. Based on the inherent relationship between these parameters, logical constraints between binary variables are defined using the Big M method. Finally, collision avoidance constraints affected by binary variables of non-fixed vehicle position relationships are obtained.

[0046]

[0047]

[0048] Among them, t gap For the minimum following distance, d min M represents the minimum distance between vehicles when stationary, and M is a constant.

[0049] Preferably, the trajectory planning problem for multi-lane vehicle platooning is defined as a mixed-integer linear programming model, including:

[0050]

[0051] Where j is the time step in the planning time domain starting from time k, and N p To plan the time domain, This represents the feasible set of vehicle states. The feasible set of inputs for vehicle decision-making and control is defined, and the vehicle's speed, acceleration trajectory, and variable acceleration inputs are restricted according to constraints. Furthermore, traffic flow continuity constraints are applied based on the preceding traffic flow status as collision avoidance constraints; f(u i (k+j),x i (k+j))≤0 is a constraint on the control quantity coupled with the vehicle's state at time k+j, specifically, the vehicle cannot make a new lane change decision before completing a lane change. For the collision avoidance constraints under the non-fixed following relationship defined above; in the objective function Let λ be the lateral lane deviation of the vehicle from the target lane at time k+j. v The weighting factor for the longitudinal velocity term is used; after linearizing the objective function and the nonlinear constraint f≤0, a MILP model is constructed to ensure the speed of online solution.

[0052] Preferably, calculating the lateral lane-changing trajectory of the vehicle at the vehicle end includes:

[0053] At the vehicle end, the lateral lane-changing trajectory is calculated using a fifth-order polynomial interpolation algorithm based on the lane-changing time window length; where:

[0054] Make the lateral position of the vehicle during the lane change process satisfy a fifth-degree polynomial equation:

[0055] q(t)=b5t 5 +b4t 4 +b3t 3 +b2t 2 +b1t+b0,

[0056] Where q(t) is the lateral position of the vehicle, b0~b5 are coefficients to be determined, and t is a time variable, with the zero moment being when the vehicle begins to change lanes.

[0057] The boundary conditions are:

[0058]

[0059] The coefficients to be determined are:

[0060] b0 = q0, b1 = b2 = 0

[0061] Where q0 represents the lateral position of the lane centerline before the lane change. and These are the lateral velocity and lateral acceleration of the vehicle at time *, q c This refers to the lateral position of the lane centerline after the lane change.

[0062] Therefore, the lateral lane-changing trajectory of the vehicle is:

[0063]

[0064] Preferably, the step of re-collecting the current state information of the vehicles and updating the state vector, updating the subset of neighboring vehicles, planning the time domain to roll forward one step, and solving the trajectory planning problem for multi-lane vehicle platooning includes:

[0065] When the current time changes from k0 to k0+1, the target lane information, the vehicle's current state information, and the overall traffic flow are fused and filtered to obtain the vehicle's current real-time longitudinal position, velocity, and acceleration trajectory, and the state vector p is updated. i (k0+1),v i (k0+1),a i (k0);

[0066] Regarding the lane-changing status of a vehicle, if the vehicle changes lanes according to the expected schedule, then as the timestamp changes from k0 to k0+1, the lane decision status n is recursively assigned based on the timestamp changes. i ((k0+1)-N c ) = n i (k0-(N c -1)),n i ((k0+1)-(N c -1))=n i (k0-(N c -2),...,n i ((k0+1)-2)=n i (k0-1), and update the most recent lane decision value based on the lane change decision at time k0.

[0067] If the vehicle does not change lanes as expected, the recursion will not proceed and the state at time k0 will remain unchanged. i ((k0+1)-N c ) = n i(k0-N c )),n i ((k0+1)-(N c -1))=n i (k0-(N c -1),...,n i ((k0+

[0068] 1)-2)=n i (k0-2), and n i ((k0+1)-1) still follows Update; now, based on the latest vehicle status, update and construct the neighbor subsets of all vehicles, starting from the current time k0+1, and resolve the planning time domain N. p Trajectory planning problem within the system.

[0069] According to another aspect of the present invention, a multi-lane platooning system for intelligent connected vehicles without a fixed configuration is provided, comprising:

[0070] The information acquisition module is used to acquire the target lane information of vehicles entering the platooning area, and to collect the current status information of the vehicles and the overall traffic flow information.

[0071] The vehicle linear dynamic model construction module establishes a vehicle linear dynamic model based on the target lane information, vehicle status information, and traffic flow status information. This model is used to describe the vehicle's longitudinal following behavior and lateral lane change decision.

[0072] The neighbor vehicle subset construction module is used to number all vehicles in the order they arrive at the platooning area. Based on the lanes that a vehicle needs to pass through to complete the platooning, a neighbor vehicle subset affected by its following and lane-changing behavior is constructed for each vehicle's linear dynamic model. For each vehicle and the vehicles in its neighbor vehicle subset, a non-fixed following relationship and collision avoidance constraint are defined to construct an optimizable configuration for multi-lane vehicle platooning.

[0073] The trajectory planning result solving module defines the trajectory planning problem of the multi-lane vehicle formation as a mixed integer linear programming model based on the optimizable configuration of the multi-lane vehicle formation and solves it to obtain the optimized trajectory planning result.

[0074] The lateral lane change trajectory calculation module is used to send the longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning result to the vehicle, execute the corresponding lateral lane change decision and longitudinal control quantity through the vehicle, and calculate the lateral lane change trajectory of the vehicle at the vehicle end.

[0075] The trajectory rolling calculation module is used to coordinate with other modules to plan and control the formation process. It collects the current state information of the vehicles and updates the state vector. It updates the subset of neighboring vehicles, rolls forward one step in the time domain and solves the trajectory planning problem of multi-lane vehicle formation. It obtains the decision control quantity at the current moment, sends it to the vehicles and executes it. This process is repeated until all vehicles enter the target lane and the formation is completed.

[0076] According to a third aspect of the present invention, a computer terminal is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can be used to perform the method described in any one of the above inventions, or to run the system described in the above inventions.

[0077] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, can be used to perform the method described in any one of the above-described inventions, or to run the system described in the above-described inventions.

[0078] By adopting the above technical solution, the present invention has at least one of the following beneficial effects compared with the prior art:

[0079] This invention provides a method and system for multi-lane platooning of intelligent connected vehicles without fixed configuration. It utilizes mixed-integer linear programming for multi-vehicle trajectory planning, centrally optimizing the lateral and longitudinal movements of all vehicles during platooning. The multi-vehicle trajectory planning model uses the longitudinal acceleration and lateral lane-changing decisions of vehicles as decision variables. The objective function considers minimizing the error between the lanes of all vehicles and the target lane and increasing the speed of all vehicles. This allows for coordinated maneuvering of vehicles, such as yielding and lane changing, enabling all vehicles in the multi-lane traffic flow within the conflict zone to quickly and efficiently enter the target lane according to the target direction of travel, thus optimizing the speed and overall traffic flow performance within the conflict zone.

[0080] This invention provides a non-fixed configuration intelligent connected vehicle multi-lane platooning method and system. It does not specifically limit the platoon configuration, but rather constructs a subset of neighboring vehicles influenced by a vehicle's following and lane-changing behaviors based on the vehicle's desired lane, thus identifying the interaction relationships between vehicles. For each vehicle and its neighboring subset, non-fixed following relationships and collision avoidance constraints are defined, providing greater freedom for the multi-vehicle trajectory planning model and further expanding the performance optimization space of the platooning process. The final platoon's vehicle order and following relationships are optimizable, exhibiting strong universality, and the platooning within the framework of uncertain configuration demonstrates high flexibility and maneuverability. Attached Figure Description

[0081] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0082] Figure 1 This is a flowchart illustrating the process of a multi-lane platooning method for intelligent connected vehicles without a fixed configuration, according to an embodiment of the present invention.

[0083] Figure 2 This is a flowchart of the upper-level centralized optimization and control of a multi-lane platooning method for intelligent connected vehicles without fixed configuration in a preferred embodiment of the present invention.

[0084] Figure 3 This is a flowchart of the process of the vehicle receiving and executing upper-level instructions in each control cycle in a preferred embodiment of the present invention.

[0085] Figure 4 This is a schematic diagram illustrating the establishment of a platooning area and the implementation of vehicle platooning in a typical three-lane scenario, as described in a specific application example of the present invention.

[0086] Figure 5 This is a schematic diagram of the vehicle-to-vehicle interaction relationship and the subset of neighboring vehicles in a specific application example of the present invention.

[0087] Figure 6 This is a comparison diagram of the passage effect in the conflict zone when no formation is performed and when formation is performed in a specific application example of the present invention; where (a) is the average time and (b) is the average passage speed.

[0088] Figure 7 This is a schematic diagram of the constituent modules of a multi-lane platooning system for intelligent connected vehicles without a fixed configuration, according to an embodiment of the present invention. Detailed Implementation

[0089] The embodiments of the present invention are described in detail below: These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

[0090] One embodiment of the present invention addresses the shortcomings of the prior art by providing a multi-lane platooning method for intelligent connected vehicles without a fixed configuration. This method centrally coordinates the lateral and longitudinal behaviors of all vehicles to carry out multi-vehicle collaborative platooning based on the desired lane selection information of vehicles passing through the conflict zone, enabling all vehicles to safely and quickly enter the target lane to form a platoon and thus efficiently pass through the conflict zone.

[0091] Specifically, such as Figure 1 As shown, the multi-lane platooning method for intelligent connected vehicles without fixed configuration provided in this embodiment may include the following operations:

[0092] S1, obtain the target lane information of vehicles entering the formation area, and collect the current status information of the vehicles and the overall traffic flow information;

[0093] S2, based on target lane information, vehicle status information and traffic flow status information, establishes a linear dynamic model of the vehicle, which is used to describe the vehicle's longitudinal following behavior and lateral lane change decision.

[0094] S3: Number all vehicles according to the order in which they arrive at the formation area. Based on the lanes that the vehicles need to pass through to complete the formation, construct a subset of neighboring vehicles that are affected by the following and lane-changing behaviors of each vehicle in the linear dynamic model of the vehicle. Define non-fixed following relationships and collision avoidance constraints for each vehicle and the vehicles in its neighboring vehicle subset to construct an optimizable configuration for multi-lane vehicle formation.

[0095] S4, based on the optimizable configuration of multi-lane vehicle formation, defines the trajectory planning problem of multi-lane vehicle formation as a mixed integer linear programming model and solves it to obtain optimized trajectory planning results;

[0096] S5 sends the longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning result to the vehicle, executes the corresponding lateral lane change decision and longitudinal control quantity through the vehicle, and calculates the lateral lane change trajectory of the vehicle at the vehicle end.

[0097] S6, collect the current state information of the vehicles again and update the state vector, update the subset of neighboring vehicles, plan the time domain forward one step and solve the trajectory planning problem of multi-lane vehicle formation, obtain the decision control quantity at the current moment, issue it to the vehicles and execute it, and repeat this process until all vehicles enter the target lane and the formation is completed.

[0098] In some preferred embodiments, S1, which involves acquiring the target lane information of vehicles entering the platooning area and collecting the current status information of the vehicles and the overall traffic flow information, may further include the following operations:

[0099] S11, detect all vehicles entering the formation area;

[0100] S12, obtain the target location and path of each vehicle and match them with the travel direction of each lane to obtain the desired lane information for the vehicle to pass through the conflict zone, i.e., the target lane information:

[0101] S13, collect the current status information of all vehicles, including the vehicle's position, speed and acceleration;

[0102] S14, Obtain overall traffic flow information, which includes: current average traffic speed and vehicle density information.

[0103] In some preferred embodiments, S2 above establishes a linear dynamic model of the vehicle based on target lane information, vehicle status information, and traffic flow status information. This model is used to describe the vehicle's longitudinal following behavior and lateral lane-changing decisions, and may further include the following operations:

[0104] A vehicle linear dynamic model is established, wherein the length of the time window provided for the vehicle to complete the lane change is determined based on the traffic flow status information and used as the parameter of the vehicle linear dynamic model, and the current state variable of the vehicle linear dynamic model is updated using the vehicle status information.

[0105] The longitudinal following behavior of the vehicle is established using an incremental second-order linear model:

[0106]

[0107] v i (k+1)=v i (k)+a i (k-1)Δt+Δa i (k)Δt,

[0108] a i (k)=a i (k-1)+Δa i (k),

[0109] Where, p i (*) and v i (*) represent the longitudinal position and velocity of vehicle i at the * discrete time, respectively, and a i (*) represents the acceleration maintained by vehicle i during the *th discrete period, Δt represents the length of the discrete time interval, and Δa i (*) indicates the change in acceleration at the *th discrete time.

[0110] The lateral lane change decision part is established using a time-delay model that describes the time taken for vehicles to change lanes and has undergone state augmentation processing:

[0111]

[0112]

[0113] Where, n i (*) indicates the decision state of vehicle i at time * regarding which lane it plans to go to. Let N represent the rightward and leftward lane change decisions made by vehicle i at time *. c This indicates the length of the time window provided for a vehicle to complete a lane change.

[0114] The state vector x of the vehicle linear dynamic model i (k) is:

[0115] x i (k)=[p i (k),v i (k),a i (k-1),n i (kN c ),n i (kN c +1),...,n i (k-1)] T

[0116] Decision input u of the vehicle linear dynamic model i (k) is:

[0117]

[0118] In some preferred embodiments, determining the length of the time window provided for the vehicle to complete the lane change in S2 above may further include the following operations:

[0119] Based on the real-time traffic flow status information, calculate the sufficient time required for vehicles to perform lane-changing maneuvers:

[0120] t c =exp(c v v flow +c d d traffic +c0)

[0121] Among them, v flow d represents the average speed of the traffic flow within the current formation area. traffic Vehicle density within the formation area; c v c d c0 and c0 are parameters used to calculate lane change time, respectively. In the offline phase, multiple linear regression is used based on experimental and collected road vehicle lane change data. We obtain, where X n Let ψ be a vector composed of the average speed of traffic flow and the vehicle density in the observation n, and let ψ be the corresponding parameter, ∈ n For the error term of the observed value n;

[0122] Online phase calculation t c And according to The length of the time window provided for the vehicle to complete the lane change is obtained.

[0123] In some preferred embodiments, the above-mentioned S3, constructing a subset of neighboring vehicles, may further include the following operations:

[0124] Construct the set of lanes that all vehicles i∈I must take to reach the target lane from their current position. or Where 'l' represents the lane number. Let i be the target lane number; for each vehicle i, obtain the subset of vehicles whose lanes intersect with its lanes on the lateral lane change. in, The elements are derived from the set I of all vehicles participating in the formation. For vehicles The set of lanes that must be taken to reach its target lane; based on the current longitudinal distance between vehicles, a subset of neighboring vehicles that may be affected by the following and lane-changing behavior of vehicle i in the prediction time domain. in, For vehicles The longitudinal position, d LIs It is the distance parameter affected by vehicle following and lane changing behavior.

[0125] In some preferred embodiments, S3 above, which defines non-fixed following relationships and collision avoidance constraints, may further include the following operations:

[0126] For vehicle i and all vehicles in the subset of vehicle i's neighboring vehicles. Define three types of auxiliary binary variables:

[0127]

[0128]

[0129]

[0130] in, and Indicate vehicle i and vehicle k at time k respectively The front-back and left-right positional relationships; Indicate vehicle i and vehicle Is the lateral distance less than one lane? This indicates the vehicle's position in the transverse lane. The lane change progress coefficient is obtained from the lateral trajectory planner of the vehicle layer. Based on the inherent relationship between these parameters, logical constraints between binary variables are defined using the Big M method. Finally, collision avoidance constraints affected by binary variables of non-fixed vehicle position relationships are obtained.

[0131]

[0132]

[0133] Among them, t gap For the minimum following distance, d min Let M be the minimum distance between vehicles in a stationary state, and M be a sufficiently large constant.

[0134] In some preferred embodiments, S4 above, which defines the trajectory planning problem of multi-lane vehicle platooning as a mixed-integer linear programming model, may further include the following operations:

[0135]

[0136] Where j is the time step in the planning time domain starting from time k, and N p To plan the time domain, This represents the feasible set of vehicle states. The feasible set of inputs for vehicle decision-making and control is defined, and the vehicle's speed, acceleration trajectory, and variable acceleration inputs are restricted according to constraints. Furthermore, traffic flow continuity constraints are applied based on the preceding traffic flow status as collision avoidance constraints; f(u i (k+j),x i (k+j))≤0 is a constraint on the control quantity coupled with the vehicle's state at time k+j, specifically, the vehicle cannot make a new lane change decision before completing a lane change. For the collision avoidance constraints under the non-fixed following relationship defined above; in the objective function Let λ be the lateral lane deviation of the vehicle from the target lane at time k+j. v The weighting factor for the longitudinal velocity term is used; after linearizing the objective function and the nonlinear constraint f≤0, a MILP model is constructed to ensure the speed of online solution.

[0137] In some preferred embodiments, the above-mentioned S5, which calculates the lateral lane-changing trajectory of the vehicle at the vehicle end, may further include the following operations:

[0138] At the vehicle end, the lateral lane-changing trajectory is calculated using a fifth-order polynomial interpolation algorithm based on the lane-changing time window length; where:

[0139] Make the lateral position of the vehicle during the lane change process satisfy a fifth-degree polynomial equation:

[0140] q(t)=b5t 5 +b4t 4 +b3t 3 +b2t 2 +b1t+b0,

[0141] Where q(t) is the lateral position of the vehicle, b0~b5 are coefficients to be determined, and t is a time variable, with the moment when the vehicle begins to change lanes being taken as the zero time t=0;

[0142] The boundary conditions are:

[0143]

[0144] The coefficients to be determined are:

[0145] b0 = q0, b1 = b2 = 0

[0146] Where q0 represents the lateral position of the lane centerline before the lane change. and These are the lateral velocity and lateral acceleration of the vehicle at time *, q c This refers to the lateral position of the lane centerline after the lane change.

[0147] Therefore, the lateral lane-changing trajectory of the vehicle is:

[0148]

[0149] In some preferred embodiments, S6 above, which involves re-collecting the current state information of the vehicle and updating the state vector, updating the subset of neighboring vehicles, planning the time domain to roll forward one step, and solving the trajectory planning problem for multi-lane vehicle platooning, may further include the following operations:

[0150] When the current time changes from k0 to k0+1, the target lane information, the vehicle's current state information, and the overall traffic flow are fused and filtered to obtain the vehicle's current real-time longitudinal position, velocity, and acceleration trajectory, and the state vector p is updated. i (k0+1),v i (k0+1),a i (k0);

[0151] Regarding the lane-changing status of a vehicle, if the vehicle changes lanes according to the expected schedule, then as the timestamp changes from k0 to k0+1, the lane decision status n is recursively assigned based on the timestamp changes. i ((k0+1)-N c ) = n i (k0-(N c -1)),n i ((k0+1)-(N c -1))=n i (k0-(N c -2),...,n i ((k0+1)-2)=n i (k0-1), and update the most recent lane decision value based on the lane change decision at time k0.

[0152] If the vehicle does not change lanes as expected, the recursion will not proceed and the state at time k0 will remain unchanged. i ((k0+1)-N c ) = n i (k0-N c )),n i ((k0+1)-(N c -1))=n i (k0-(N c -1),...,n i ((k0+

[0153] 1)-2)=n i (k0-2), and n i ((k0+1)-1) still follows Update; now, based on the latest vehicle status, update and construct the neighbor subsets of all vehicles, starting from the current time k0+1, and resolve the planning time domain N. p Trajectory planning problem within the system.

[0154] The technical solution provided by the above embodiments of the present invention will be further described in detail below with reference to a specific application example and accompanying drawings.

[0155] like Figure 2 and Figure 3 As shown, this specific application example provides a multi-lane platooning method for intelligent connected vehicles without fixed configuration, by using a... Figure 4 The collision zone upstream of a typical three-lane road junction, shown as a specific case, is used in an experiment conducted in the SUMO simulation environment to illustrate the multi-lane formation method of the present invention, specifically including the following steps:

[0156] S1, the roadside unit obtains the target lane information of vehicles entering the formation area through dedicated short-range communication, and collects the current status information of vehicles and the overall traffic flow through roadside and on-board sensors;

[0157] S2, establish a linear dynamic model of the vehicle, considering the vehicle's longitudinal following behavior and lateral lane change decision, determine the length of the time window provided for the vehicle to complete the lane change based on the traffic flow conditions as the model parameter, and define the state variables of the model at the current moment using the detected vehicle state information.

[0158] S3: Number all vehicles according to the order in which they arrive at the formation area. For each vehicle, construct a subset of neighboring vehicles affected by its following and lane-changing behaviors. Define non-fixed following relationships and collision avoidance constraints for each vehicle and the vehicles in its neighboring vehicle subset.

[0159] S4 defines the trajectory planning problem of multi-lane vehicle platooning as a mixed integer linear programming model and solves it. The decision variables are the longitudinal variable acceleration of the vehicle and the lateral decision to change lanes to the left or right for each control cycle. The objective function considers controlling the vehicle to change lanes as soon as possible to join the target platoon and to increase the longitudinal speed of the vehicle during the platooning process.

[0160] S5 sends the longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimization results to the vehicle through dedicated short-range communication. The vehicle executes the lateral lane change decision and longitudinal control quantity at the current moment. The lateral lane change trajectory of the vehicle is calculated at the vehicle end using a fifth-order polynomial interpolation algorithm based on the lane change time window length.

[0161] S6, at the next moment, collect vehicle information again and update the state vector, update the vehicle's neighbor subset, plan the time domain forward one step and solve the multi-vehicle trajectory planning problem, and let the vehicle execute the decision control quantity of the current moment obtained by optimization.

[0162] Specifically:

[0163] S1: Obtain the target lane information of vehicles entering the platooning area, and collect the current status information of the vehicles and the overall traffic flow situation, as detailed below:

[0164] S1.1, Detect all vehicles entering the formation area;

[0165] S1.2, Obtain the desired lane information for vehicles entering the platooning area to pass through the conflict zone:

[0166] S1.3 collects status information of all vehicles, including position, speed and acceleration, through roadside and on-board sensors;

[0167] S1.4, obtain information such as the current average traffic speed and vehicle density;

[0168] S2. Establish a linear dynamic model of the vehicle, considering its longitudinal following behavior and lateral lane-changing decisions. The longitudinal kinematics part adopts an incremental second-order linear model:

[0169]

[0170] v i (k+1)=v i (k)+a i (k-1)Δt+Δa i (k)Δt,

[0171] a i (k)=a i (k-1)+Δa i (k),

[0172] Where, p i (k) and v i (k) represents the longitudinal position and velocity of vehicle i at the k-th discrete time, a i (k-1) represents the acceleration maintained by vehicle i during the (k-1)th discrete period; Δt represents the length of the discrete time interval; Δa i (k) represents the change in acceleration at the k-th discrete time. The lateral kinematics part uses a time-delay model that describes the vehicle's lane change time and has undergone state augmentation:

[0173]

[0174]

[0175] in Indicates that vehicle i is in the... The decision-making state of constantly planning which lane to take. N represents the decision made by vehicle i at time k to change lanes to the right or left. c This represents the length of the time window provided for the vehicle to complete the lane change. The state vector of the vehicle model is x. i (k)=[p i (k),v i (k),a i (k-1),n i (kN c ),n i (kN c +1),...,n i (k-1)] T The control (decision) input is

[0176] It is also necessary to calculate, based on real-time traffic flow and conditions, the sufficient time required for vehicles to perform lane-changing maneuvers according to the lane-change time model: t c =exp(c v v flow +c d d traffic +c0), where v flow d represents the average speed of the traffic flow within the current formation area. traffic Vehicle density within the formation area; c v c d c0 are the parameters of the lane change time model. In the offline phase, the model is constructed using multiple linear regression based on experimental and collected road vehicle lane change data. We obtain X n Let ψ be a vector composed of the average speed of traffic flow and the vehicle density in the observation n, and let ψ be the corresponding parameter, ∈n The error term for the observed value n; t is calculated in the online phase. c And according to The length of the time window provided for the vehicle to complete the lane change is obtained.

[0177] Finally, based on the detected current vehicle state and the calculated lane-changing time window length N, c It allows you to define the state variables of the vehicle model and their initial values ​​at the current moment.

[0178] S3: Number all vehicles according to their arrival order in the platooning area, and construct vehicle set I. For example... Figure 5 As shown, the interaction relationships between vehicles in the workshop are analyzed. A set of lanes that all vehicles i∈I must take to reach the target lane from their current position is constructed. or in Let i be the target lane for vehicle i; for each vehicle i, obtain the subset of vehicles whose lanes intersect with its lateral movement. Based on the current longitudinal distance between vehicles, a subset of neighboring vehicles that may be affected by vehicle i's following and lane-changing behavior in the prediction time domain is identified. Where d LIS It is the distance parameter affected by vehicle following and lane changing behavior.

[0179] For vehicle i and all vehicles in the subset of vehicle i's neighboring vehicles. To express the non-fixed following relationship, three types of auxiliary binary variables are defined. in and Indicate vehicle i and vehicle k at time k respectively The front-back and left-right positional relationships; Indicate vehicle i and vehicle Is the lateral distance less than one lane? This indicates the vehicle's position in the transverse lane. This refers to the lane change progress coefficient obtained from the lateral trajectory planner at the vehicle layer. Based on the inherent relationships among these parameters, logical constraints between binary variables are defined using the Big M method, specifically expressed as follows:

[0180]

[0181]

[0182]

[0183]

[0184] Finally, we can obtain the collision avoidance constraints affected by the binary variable of non-fixed workshop position relationships:

[0185]

[0186]

[0187] Among them, t gap For the minimum following distance, d min Let M be the minimum distance between vehicles in a stationary state, and M be a sufficiently large constant.

[0188] S4 defines the trajectory planning problem for multi-lane vehicle platooning as a mixed-integer linear programming model:

[0189]

[0190] The decision variables are the longitudinal acceleration of all vehicles and the decision to change lanes to the left or right for each control cycle; the objective function considers controlling vehicles to execute lane changes as quickly as possible to join the target platoon and to increase the longitudinal speed of vehicles during platooning. Let λ be the lateral lane deviation of the vehicle from the target lane at time k+j. v The weighting factor for the velocity term; A and B are the system matrix and control matrix derived from the established system model, respectively; N p For planning the time domain; This represents the feasible set of vehicle states. The feasible set serves as the input for vehicle decision-making and control, limiting the vehicle's speed v based on traffic rules, vehicle dynamics constraints, and ride comfort. min ≤v i (k+j)≤v max acceleration a min ≤a i (k+j)≤a max and variable acceleration input -Δa max ≤Δa i (k+j)≤Δa max The range is defined, and traffic continuity (collision avoidance) constraints are applied based on the traffic flow conditions ahead: if f(u i (k+j),x i (k+j))≤0 is a constraint on the control quantity coupled with the vehicle's state at time k+j, specifically, the vehicle cannot make a new lane change decision before completing a lane change. The collision avoidance constraints under the non-fixed following relationship defined above;

[0191] Linearize the nonlinear terms containing absolute values ​​in the objective function: And introduce additional constraints m i (k+j)≥0, m′ i (k+j)≥0. For nonlinear constraints f(u) i (k+j),x i Linearize (k+j)≤0 to obtain the linearized result. After linearization, the multi-vehicle trajectory planning model becomes a MILP model, which helps to solve it quickly online.

[0192] S5, after solving the trajectory planning optimization problem, distributes the longitudinal position, velocity, acceleration trajectory, and lane change decision for each vehicle from the optimization results to the vehicles. The execution content on the vehicle side is as follows: Figure 3 As shown, the vehicle makes a lateral lane change decision at the current moment. or With longitudinal control quantity Δa i (k0) enables the vehicle to track the optimized trajectory p in the next time step. i (k0+1), v i (k0+1). The lateral lane-change trajectory of the vehicle is calculated at the vehicle end using a fifth-order polynomial interpolation algorithm based on the lane-change time window length, ensuring that the lateral position of the vehicle during the lane-change process satisfies the fifth-order polynomial equation: q(t)=b5t 5 +b4t 4 +b3t 3 +b2t 2 +b1t+b0, and according to the boundary conditions The coefficients to be determined are b0 = q0, b1 = b2 = 0. Where q0 represents the lateral position of the lane centerline before the lane change, q c This represents the lateral position of the lane centerline after the lane change. This gives the vehicle's lateral trajectory after the lane change.

[0193] S6, at the next moment k0→k0+1, collect vehicle information again and update the state vector to obtain the vehicle's current real-time longitudinal position, velocity, and acceleration, and update the state variable p. i (k0+1),v i (k0+1),a i (k0); For the lane change status of the vehicle, if the vehicle changes lanes according to the expected schedule, then as the timestamp changes from k0 to k0+1, the lane decision status n is recursively assigned based on the timestamp changes. i ((k0+1)-N c ) = n i (k0-(N c -1)),ni ((k0+1)-(N c -1))=n i (k0-(N c -2),...,n i ((k0+1)-2)=n i (k0-1), and update the most recent lane decision value based on the lane change decision at time k0. If a vehicle fails to change lanes as expected due to unexpected factors such as crosswinds, the recursion will not proceed, n i ((k0+1)-N c ) = n i (k0-N c )),n i ((k0+1)-(N c -1))=n i (k0-(N c -1),...,n i ((k0+

[0194] 1)-2)=n i (k0-2), these lane decision values ​​remain unchanged relative to k0 at time k0+1, while n i ((k0+1)-1) still follows renew.

[0195] Next, based on the latest vehicle status, update the vehicle's neighbor subset, and also roll the planning time domain forward one step to resolve the multi-vehicle trajectory planning problem, and let the vehicle execute the decision control quantity of the first time step obtained by the optimization solution;

[0196] Repeat the above process, performing rolling time-domain optimization and control, until all vehicles enter the target lane and platooning is complete. To analyze the effectiveness of platooning, experiments were conducted under different traffic flow conditions. The traffic performance of vehicles autonomously following and changing lanes to enter the target lane within the conflict zone without platooning control was compared with the traffic performance of vehicles in platooning. The results are as follows: Figure 6 As shown in (a) and (b), platooning significantly improves both the time it takes for vehicles to enter the target lane and the speed of traffic passing through the conflict zone. Specifically, under heavy traffic conditions (over 1500 vehicles / hour / lane), congestion occurred in the conflict zone without platooning; however, with platooning, the average time for vehicles to enter the target lane was reduced by up to 84.4%, and the average speed of vehicles passing through the conflict zone increased by up to 58.6%.

[0197] One embodiment of the present invention provides a multi-lane platooning system for intelligent connected vehicles without a fixed configuration.

[0198] Specifically, such as Figure 7 As shown, the intelligent connected vehicle multi-lane platooning system without a fixed configuration provided in this embodiment may further include the following modules:

[0199] The information acquisition module is used to acquire the target lane information of vehicles entering the platooning area, and to collect the current status information of the vehicles and the overall traffic flow information.

[0200] The vehicle linear dynamic model construction module establishes a vehicle linear dynamic model based on target lane information, vehicle status information, and traffic flow status information. This model is used to describe the vehicle's longitudinal following behavior and lateral lane change decision.

[0201] The neighbor vehicle subset construction module is used to number all vehicles in the order they arrive at the platooning area. Based on the lanes that a vehicle needs to pass through to complete the platooning, a neighbor vehicle subset affected by its following and lane-changing behavior is constructed for each vehicle's linear dynamic model. For each vehicle and the vehicles in its neighbor vehicle subset, a non-fixed following relationship and collision avoidance constraint are defined to obtain multi-lane vehicle platooning.

[0202] The trajectory planning result solving module defines the trajectory planning problem of multi-lane vehicle formation as a mixed integer linear programming model based on the optimizable configuration of multi-lane vehicle formation and solves it to obtain the optimized trajectory planning result.

[0203] The lateral lane change trajectory calculation module is used to send the longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning results to the vehicle, and the vehicle executes the corresponding lateral lane change decision and longitudinal control quantity, and calculates the lateral lane change trajectory of the vehicle at the vehicle end.

[0204] The trajectory rolling calculation module is used to coordinate with other modules to plan and control the formation process. It collects the current state information of the vehicles and updates the state vector. It updates the subset of neighboring vehicles, rolls forward one step in the time domain and solves the trajectory planning problem of multi-lane vehicle formation. It obtains the decision control quantity at the current moment, sends it to the vehicles and executes it. This process is repeated until all vehicles enter the target lane and the formation is completed.

[0205] It should be noted that the steps in the method provided by the present invention can be implemented using corresponding modules, devices, units, etc. in the system. Those skilled in the art can refer to the technical solution of the method to realize the composition of the system. That is, the embodiments in the method can be understood as preferred examples for building the system, and will not be elaborated here.

[0206] One embodiment of the present invention provides a computer terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can be used to perform any of the methods in the above embodiments of the present invention, or to run any of the systems in the above embodiments of the present invention.

[0207] Optionally, the memory is used to store programs; the memory may include volatile memory, such as random-access memory (RAM), such as static random-access memory (SRAM), double data rate synchronous dynamic random-access memory (DDR SDRAM), etc.; the memory may also include non-volatile memory, such as flash memory. The memory is used to store computer programs (such as application programs, functional modules, etc. that implement the above methods), computer instructions, etc., and the aforementioned computer programs, computer instructions, etc., can be partitioned and stored in one or more memories. Furthermore, the aforementioned computer programs, computer instructions, data, etc., can be accessed by the processor.

[0208] The aforementioned computer programs, computer instructions, etc., can be stored in partitions within one or more memory locations. Furthermore, the aforementioned computer programs, computer instructions, data, etc., can be accessed by a processor.

[0209] A processor is used to execute computer programs stored in memory to implement the various steps of the methods or various modules of the systems involved in the above embodiments. For details, please refer to the relevant descriptions in the preceding method and system embodiments.

[0210] The processor and memory can be separate structures or integrated structures. When the processor and memory are separate structures, they can be coupled together via a bus.

[0211] One embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program can be used to perform the method of any of the above embodiments of the present invention, or to run the system of any of the above embodiments of the present invention.

[0212] The multi-lane platooning method and system for intelligent connected vehicles without fixed configuration provided in the above embodiments of the present invention do not pre-determine a fixed platoon configuration. Instead, the platooning is controlled to flexibly enter the target lane to form a platoon based on the desired lane selection information when vehicles pass through the conflict zone. By analyzing the interaction and influence relationships between vehicles, a subset of neighboring vehicles is constructed. Based on the subset of neighboring vehicles, non-fixed following relationships and collision avoidance constraints are defined to achieve a flexible and variable platoon configuration. The multi-vehicle trajectory planning problem in the platooning process is defined as an optimization model of mixed integer linear programming. The decision variables simultaneously consider the lateral and longitudinal movements of vehicles, and the objective function optimizes the lane deviation and driving speed that change over time. A rolling time-domain optimization and control framework is used for platooning. The vehicle state is updated at each sampling time, the multi-vehicle trajectory planning problem in the planning time domain is solved, and the decision results and control variables at the current time are executed. The present invention effectively improves the flexibility, maneuverability, and platooning efficiency of platooning in traffic flow and increases the traffic flow speed in the conflict zone.

[0213] Any matters not covered in the above embodiments of the present invention are well-known in the art.

[0214] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. A method for multi-lane platooning of intelligent connected vehicles without fixed configuration, characterized in that, include: Obtain the target lane information of vehicles entering the platooning area, and collect the current status information of the vehicles and the overall traffic flow information; Based on the target lane information, vehicle status information, and traffic flow status information, a linear dynamic model of the vehicle is established. This model is used to describe the vehicle's longitudinal following behavior and lateral lane change decision. All vehicles are numbered according to the order in which they arrive at the platooning area. Based on the lanes that a vehicle needs to pass through to complete the platooning, a subset of neighboring vehicles affected by its following and lane-changing behaviors is constructed for each vehicle's linear dynamic model. For each vehicle and the vehicles in its neighboring vehicle subset, non-fixed following relationships and collision avoidance constraints are defined to construct an optimizable configuration for multi-lane vehicle platooning. Based on the optimizable configuration of the multi-lane vehicle formation, the trajectory planning problem of the multi-lane vehicle formation is defined as a mixed integer linear programming model and solved to obtain the optimized trajectory planning result. The longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning result are sent to the corresponding vehicle. The vehicle executes the corresponding lateral lane change decision and longitudinal control quantity, and the lateral lane change trajectory of the vehicle is calculated at the vehicle end. Collect the current state information of the vehicles and update the state vector. Update the subset of neighboring vehicles of each vehicle. Scroll forward one step in the time domain and solve the trajectory planning problem of multi-lane vehicle formation. Obtain the decision control quantity at the current moment, issue it to the vehicles and execute it. Repeat this process until all vehicles enter the target lane and the formation is completed. in: Construct a subset of neighboring vehicles, including: Build all vehicles The set of lanes that must be taken from the current location to the target lane. ,in, Lane numbering, For vehicles The target lane number, This represents the length of the time window provided for a vehicle to complete a lane change; for each vehicle This yields a subset of vehicles that intersect with its lanes during lateral lane changes. ,in, For the collection of all vehicles participating in the formation elements, For vehicles The set of lanes that must be taken to reach its target lane; based on the current longitudinal distance between vehicles, identify the lanes that may be affected by this vehicle in the prediction time domain. The subset of neighboring vehicles affected by following and lane-changing behavior ,in, For vehicles The vertical position, The distance parameter is affected by vehicle following and lane changing behaviors; The defined non-fixed following relationships and collision avoidance constraints include: For vehicles and all in the vehicle The neighbor's vehicle subgroup of vehicles Define three types of auxiliary binary variables: in, and Instructions respectively Time vehicle With vehicles The front-back and left-right positional relationships; Instructing vehicles With vehicles Is the lateral distance less than one lane? Indicates vehicle The location of the cross lane. The lane change progress coefficient is obtained from the lateral trajectory planner of the vehicle layer. Based on the inherent relationship between these parameters, logical constraints between binary variables are defined using the Big M method. Finally, collision avoidance constraints affected by binary variables of non-fixed vehicle position relationships are obtained. , , in, and Representing vehicles In the The longitudinal position and velocity at discrete moments, Minimum following distance, This represents the minimum distance between vehicles when stationary. It is a constant.

2. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 1, characterized in that, The process of acquiring the target lane information of vehicles entering the platooning area and collecting the current status information of the vehicles and the overall traffic flow information includes: Inspect all vehicles entering the platoon area; Obtain the target location and path of each vehicle and match them with the travel direction of each lane to obtain the desired lane information for the vehicle to pass through the conflict zone, i.e., the target lane information: Collect current status information of all vehicles, including vehicle position, speed, and acceleration; Obtain overall traffic flow information, which includes current average traffic speed and vehicle density information.

3. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 1, characterized in that, Based on the target lane information, vehicle status information, and traffic flow status information, a linear dynamic model of the vehicle is established. This model is used to describe the vehicle's longitudinal following behavior and lateral lane-changing decisions, including: A vehicle linear dynamic model is established, wherein the length of the time window provided for a vehicle to complete a lane change is determined based on the traffic flow status information and used as a parameter of the vehicle linear dynamic model, and the current time state variable of the vehicle linear dynamic model is updated using the vehicle status information. The longitudinal following behavior of the vehicle is established using an incremental second-order linear model: , , , in, and Representing vehicles In the The longitudinal position and velocity at discrete moments, Indicates vehicle In the The acceleration maintained over a discrete period, Indicates the length of the discrete time interval. Indicates the first The change in acceleration at discrete moments; The lateral lane change decision part is established using a time-delay model that describes the time taken for vehicles to change lanes and has undergone state augmentation processing: , , in, Indicates vehicle In the The decision-making state of constantly planning which lane to take. , Representing vehicles In the The decision to change lanes to the right or left that must be made at any time; The state vector of the vehicle linear dynamic model for: The decision input of the vehicle linear dynamic model for: 。 4. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 3, characterized in that, The time window length for determining the completion of a lane change by a vehicle includes: Based on the real-time traffic flow status information, calculate the sufficient time required for vehicles to perform lane-changing maneuvers: in, The average speed of the traffic flow within the current formation area. Vehicle density within the formation area; , and These are the parameters used to calculate lane change time. In the offline phase, these parameters are calculated using multiple linear regression based on experimental and collected road vehicle lane change data. We obtained, among which, For observation A vector composed of average traffic speed and vehicle density in the medium flow. For the corresponding parameters, For observations Error term; Online phase computing And according to The length of the time window provided for the vehicle to complete the lane change is obtained.

5. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 4, characterized in that, The trajectory planning problem for multi-lane vehicle platooning is defined as a mixed-integer linear programming model, including: in, From The time step within the planning time domain that begins at a given moment. To plan the time domain, This represents the feasible set of vehicle states. and These are the system matrix and control matrix derived from the established system model, respectively. The feasible set of inputs for vehicle decision-making and control is defined by constraints on vehicle speed, acceleration trajectory and variable acceleration input, and traffic flow continuity constraints are applied based on the traffic flow status ahead as collision avoidance constraints. To couple the vehicle in the first The constraint of the control quantity on the state at any moment is that the vehicle cannot make a new lane change decision before completing a lane change; For the collision avoidance constraints under the non-fixed following relationship defined above; in the objective function For the vehicle in the Lateral lane deviation from the target lane at any given time. The weighting factor for the longitudinal velocity term; for the objective function and nonlinear constraints After linearization, a MILP model is constructed to ensure online solution speed.

6. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 5, characterized in that, The calculation of the vehicle's lateral lane-changing trajectory at the vehicle end includes: At the vehicle end, the lateral lane-changing trajectory is calculated using a fifth-order polynomial interpolation algorithm based on the lane-changing time window length; where: Make the lateral position of the vehicle during the lane change process satisfy a fifth-degree polynomial equation: + + + + ， in, The vehicle's lateral position. The coefficients to be determined are: The time variable is used, with the moment when the vehicle begins to change lanes as time zero. =0; The boundary conditions are: , The coefficients to be determined are: ， ， ， ， in, The lateral position of the center line of the lane before changing lanes. and These represent the vehicle's lateral velocity and lateral acceleration at time *. This refers to the lateral position of the lane centerline after the lane change. Therefore, the lateral lane-changing trajectory of the vehicle is: 。 7. The method for multi-lane platooning of intelligent connected vehicles without fixed configuration according to claim 6, characterized in that, The process of re-collecting the current state information of the vehicles and updating the state vector, updating the subset of neighboring vehicles, planning the trajectory planning problem of multi-lane vehicle platooning by rolling forward one step in the time domain, includes: Current moment by Become At the same time, the target lane information, the current state information of the vehicle, and the overall traffic flow are fused and filtered to obtain the vehicle's current real-time longitudinal position, speed, and acceleration trajectory, and the state vector is updated. ; Regarding the lane-changing status of a vehicle, if the vehicle changes lanes as expected, the time stamp will change accordingly. The lane decision status is recursively assigned based on the changes in timestamps. and according to The lane change decision is updated with the most recent lane decision value at any given moment. ; If the vehicle does not change lanes as expected, the process will not be repeated and will remain unchanged. The state at any given moment, ,and Still in accordance with Update; at this point, based on the latest vehicle status, update and construct a subset of neighbors for all vehicles, using the current time. Starting with the time domain of the planning algorithm, we re-solve the problem. Trajectory planning problem within the system.

8. A multi-lane platooning system for intelligent connected vehicles without a fixed configuration, characterized in that, include: The information acquisition module is used to acquire the target lane information of vehicles entering the platooning area, and to collect the current status information of the vehicles and the overall traffic flow information. The vehicle linear dynamic model construction module establishes a vehicle linear dynamic model based on the target lane information, vehicle status information, and traffic flow status information. This model is used to describe the vehicle's longitudinal following behavior and lateral lane change decision. The neighbor vehicle subset construction module is used to number all vehicles in the order they arrive at the platooning area. Based on the lanes that a vehicle needs to pass through to complete the platooning, a neighbor vehicle subset affected by its following and lane-changing behavior is constructed for each vehicle's linear dynamic model. For each vehicle and the vehicles in its neighbor vehicle subset, a non-fixed following relationship and collision avoidance constraint are defined to construct an optimizable configuration for multi-lane vehicle platooning. The trajectory planning result solving module defines the trajectory planning problem of the multi-lane vehicle formation as a mixed integer linear programming model based on the optimizable configuration of the multi-lane vehicle formation and solves it to obtain the optimized trajectory planning result. The lateral lane change trajectory calculation module is used to send the longitudinal position, speed, acceleration trajectory and lane change decision of each vehicle in the optimized trajectory planning result to the vehicle, execute the corresponding lateral lane change decision and longitudinal control quantity through the vehicle, and calculate the lateral lane change trajectory of the vehicle at the vehicle end. The trajectory rolling calculation module is used to coordinate with other modules to plan and control the formation process. It collects the current state information of the vehicles and updates the state vector. It updates the subset of neighboring vehicles, rolls forward one step in the time domain and solves the trajectory planning problem of multi-lane vehicle formation. It obtains the decision control quantity at the current moment, sends it to the vehicles and executes it. This process is repeated until all vehicles enter the target lane and the formation is completed. in: The neighbor vehicle subset construction module constructs a neighbor vehicle subset, including: Build all vehicles The set of lanes that must be taken from the current location to the target lane. ,in, Lane numbering, For vehicles The target lane number, This represents the length of the time window provided for a vehicle to complete a lane change; for each vehicle This yields a subset of vehicles that intersect with its lanes during lateral lane changes. ,in, For the collection of all vehicles participating in the formation elements, For vehicles The set of lanes that must be taken to reach its target lane; based on the current longitudinal distance between vehicles, identify the lanes that may be affected by this vehicle in the prediction time domain. The subset of neighboring vehicles affected by following and lane-changing behavior ,in, For vehicles In the The vertical position at discrete moments. The distance parameter is affected by vehicle following and lane changing behaviors; The neighbor vehicle subset construction module defines non-fixed following relationships and collision avoidance constraints, including: For vehicles and all in the vehicle The neighbor's vehicle subgroup of vehicles Define three types of auxiliary binary variables: in, and Instructions respectively Time vehicle With vehicles The front-back and left-right positional relationships; Instructing vehicles With vehicles Is the lateral distance less than one lane? Indicates vehicle The location of the cross lane. For vehicles The location of the cross lane. The lane change progress coefficient is obtained from the lateral trajectory planner of the vehicle layer. Based on the inherent relationship between these parameters, logical constraints between binary variables are defined using the Big M method. Finally, collision avoidance constraints affected by binary variables of non-fixed vehicle position relationships are obtained. , , in, and Representing vehicles In the The longitudinal position and velocity at discrete moments, Minimum following distance, This represents the minimum distance between vehicles when stationary. It is a constant.

9. A computer terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it can be used to perform the method of any one of claims 1-7, or to run the system of claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program can be used to perform the method of any one of claims 1-7, or to run the system of claim 8.

Images