Lane-changing multi-lane intersection networked vehicle dispatching method and system

Abstract

The application discloses a multi-lane signal-free intersection networked vehicle scheduling method and system considering lane changing, and relates to the technical field of intelligent traffic and automatic driving vehicle control. In the spatial dimension, the initial randomness of the traffic flow is absorbed by the lane changing area; in the time dimension, the maximum parallel passing potential of the intersection is squeezed by the following area, greatly optimizing the overall evacuation efficiency of the intersection. Through active lane changing cooperation in the far end, the extreme imbalance of the flow distribution between lanes is eliminated, effectively avoiding the single-lane long queuing phenomenon caused by random distribution of vehicles. In the micro level, based on the optimal control model, the vehicle is planned with an energy-optimal smooth trajectory, so that the energy loss caused by frequent acceleration and deceleration is reduced to the maximum extent.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation and autonomous vehicle control technology, and in particular to a method and system for scheduling connected vehicles at multi-lane unsignalized intersections that takes lane changes into account. Background Technology

[0002] In complex urban road networks, at-grade intersections, as nodes for traffic convergence, turning, and dispersal, are bottlenecks in road capacity. Due to spatial competition and trajectory conflicts arising from traffic flows in different directions, intersections often become hotspots for traffic congestion and accident clusters. Related research indicates that the traffic efficiency of intersection areas directly determines the overall operational level of the road network, and a significant proportion of traffic accidents are related to intersections. Therefore, optimizing traffic control strategies at intersections is of paramount importance for improving overall traffic safety and efficiency.

[0003] For a long time, traffic lights have been the primary means of allocating right-of-way at intersections. While signal control has played a crucial role in separating conflicting traffic flows and ensuring basic safety, it is essentially a passive control method based on "spatiotemporal barriers." It suffers from significant efficiency losses: even advanced adaptive signal control struggles to completely eliminate wasted green light intervals and queue overflow effects. Frequent starts and stops by vehicles at red lights not only reduce traffic capacity but also significantly increase fuel consumption and pollutant emissions. Faced with increasingly complex mixed traffic environments, traditional signal timing schemes often fail to consider the dynamic characteristics of different types of vehicles, easily leading to a waste of spatiotemporal resources at intersections.

[0004] Currently, most existing intersection scheduling research is limited to the assumption of lane changing being prohibited, that is, assuming that connected vehicles (CAVs) are already in their target lanes when entering the controlled area, and only considering the longitudinal trajectory planning of the vehicles. However, in real-world complex traffic environments, CAVs often approach from random lanes and have different target directions (such as left turn, straight ahead, right turn), thus requiring lane changes. Existing single-vehicle lane-changing algorithms are prone to path deadlock when multiple vehicles change lanes concurrently, and lack consideration for joint optimization of lane changing and intersection scheduling. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes a method and system for scheduling connected vehicles at multi-lane unsignalized intersections that considers lane changes. This invention adopts a two-stage collaborative framework to decouple complex intersection traffic into lateral allocation of the lane-changing zone and longitudinal scheduling of the following zone, aiming to completely solve the deadlock problem of multi-vehicle lane changes and achieve globally optimal vehicle scheduling at the intersection with extremely low computational burden.

[0006] The technical solution of this invention is as follows:

[0007] On the one hand, the present invention provides a method for scheduling connected vehicles at multi-lane unsignalized intersections that considers lane changes, comprising the following steps:

[0008] Determine the upstream road segment area of ​​the unsignalized intersection and construct a relative coordinate system for the upstream road segment area of ​​the unsignalized intersection. Spatially divide the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions: the lane-changing area and the following area.

[0009] In the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, the road is divided into discrete grid nodes. The actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area are mapped to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection to obtain the initial relative coordinates of each intelligent connected vehicle and determine the target lane of each intelligent connected vehicle.

[0010] The target lane assignment is modeled as a 0-1 integer assignment planning problem. The 0-1 integer assignment planning problem is solved iteratively by minimizing the physical cost of lane changing for the entire fleet. The assignment scheme and its corresponding physical cost for each iteration are obtained. The set of collision-free paths for the assignment schemes is solved in ascending order of physical cost of lane changing. Finally, the set of collision-free paths with the minimum physical cost of lane changing is obtained, including the collision-free paths of each intelligent connected vehicle.

[0011] According to the collision-free path set, the intelligent connected vehicles change lanes in the lane change zone and enter the following zone, and plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times.

[0012] Based on the optimal acceleration trajectory planned at the micro level for each intelligent connected vehicle, the intelligent connected vehicle travels to the parking line and uses the graph relationship method to achieve conflict-free passage in the scheduling area.

[0013] Furthermore, the process of determining the upstream road segment area of ​​the unsignalized intersection and constructing a relative coordinate system for the upstream road segment area of ​​the unsignalized intersection spatially divides the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions: a lane-changing area and a following area. Specifically:

[0014] First, the central area within the stop line of the unsignalized intersection is designated as the dispatch zone, and the length of the upstream road segment of the unsignalized intersection is set as follows: The distance from the parking line outside the dispatch area is The area upstream of the unsignalized intersection is spatially divided into a lane-changing zone and a following zone, with lengths of [missing information]. and And satisfy The lane-changing zone is located on the outermost edge of the upstream section of the unsignalized intersection. Intelligent connected vehicles change lanes in the lane-changing zone and enter the target lane. The following zone is the area adjacent to the dispatch zone. Intelligent connected vehicles are prohibited from changing lanes and must follow the intelligent connected vehicle in front in the designated target lane.

[0015] Take the point located on the innermost lane, on the boundary line between the lane-changing area and the following lane, as the origin, and take the direction away from the intersection as... In the axial direction, from the innermost lane to the outermost lane. Along the axis, a relative coordinate system is constructed for the upstream road segment area of ​​the unsignalized intersection.

[0016] Furthermore, in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, the road is divided into discrete grid nodes. The actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area are mapped to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, thus obtaining the initial relative coordinates of each intelligent connected vehicle. At the same time, the target lane of each intelligent connected vehicle is determined, specifically as follows:

[0017] Setting up intelligent connected vehicles The actual two-dimensional physical coordinates are The mapped coordinates of the upstream road segment of an unsignalized intersection in the relative coordinate system are obtained by solving the following problem of minimizing Euclidean distance. :

[0018] (1);

[0019] in, The set of natural numbers, A relative coordinate system representing the upstream road segment area of ​​an unsignalized intersection;

[0020] At the same time, at the boundary of the lane-changing zone, there are also a number of target location They are to be allocated to various vehicles. The number of vehicles.

[0021] Furthermore, the target lane allocation is modeled as a 0-1 integer allocation planning problem. This problem is solved iteratively by minimizing the physical cost of lane changes for the entire platoon, yielding allocation schemes and their corresponding physical costs for each iteration. The set of collision-free paths for each allocation scheme is then calculated sequentially from smallest to largest physical cost, ultimately resulting in the set of collision-free paths with the minimum actual physical cost of lane changes. This process includes the following steps:

[0022] A1: Construct the lane-changing physical cost matrix, preference matrix, and allocation matrix;

[0023] The lane-changing physical cost matrix is ​​as follows:

[0024] (2);

[0025] (3);

[0026] in, For the physical cost matrix of lane changing, For the lane-changing physical cost matrix, the first... Line number The elements of the column represent intelligent connected vehicles. Assigned to target location The cost of allocation The target location number. Represents the set of positive integers;

[0027] Based on the target lane of each intelligent connected vehicle, the preferences of the intelligent connected vehicles are defined by a preference matrix, expressed as follows:

[0028] (5);

[0029] (6);

[0030] in, For the preference matrix, The first in the preference matrix Line number The column elements represent intelligent connected vehicles taking lane preference into account. Can it be assigned to the target location? ;

[0031] The allocation matrix is ​​represented as follows:

[0032] (8);

[0033] (9);

[0034] in, For the allocation matrix, For the allocation matrix, the first Line number The elements of the column represent intelligent connected vehicles. Whether successfully assigned to the target location ;

[0035] Allocation matrix There is a necessary constraint: each column and each row can have only one element that is 1;

[0036] A2: Based on the lane-switching physical cost matrix, preference matrix, and allocation matrix, construct a 0-1 integer programming allocation problem and solve it iteratively to obtain the allocation scheme and its corresponding lane-switching physical cost for each iteration;

[0037] The 0-1 integer programming allocation problem is as follows:

[0038] (10);

[0039] Solve the 0-1 integer programming assignment problem to obtain the assignment matrix. ;

[0040] A3: Based on the allocation scheme of each iteration and its corresponding lane-changing physical cost, solve for the set of collision-free paths with the minimum actual lane-changing physical cost.

[0041] Furthermore, A3 specifically includes the following steps:

[0042] A3.1: Setting the initial iteration index =1;

[0043] A3.2: Obtain the allocation scheme with the minimum lane-changing physical cost as the optimal allocation scheme, perform path planning, and solve for its collision-free path set and actual lane-changing physical cost;

[0044] In the relative coordinate system of the upstream section of an unsignalized intersection, intelligent connected vehicles are restricted to moving only in orthogonal directions. During concurrent lane changes, intelligent connected vehicles face the following three micro-conflicts:

[0045] (1) Node conflict: Intelligent connected vehicles attempt to occupy the same discrete grid node as their target location at the same time step;

[0046] (2) Edge conflict: Two intelligent connected vehicles exchange with each other in the same time step between adjacent discrete grid nodes;

[0047] (3) Intermediate conflict: Considering the physical size of intelligent connected vehicles and the time difference of lane changing, an intelligent connected vehicle forcibly cuts in before another intelligent connected vehicle has completely left the discrete grid node;

[0048] The collision-free path planning process is as follows:

[0049] For the The current allocation scheme for the next iteration ,definition For the corresponding set of collision-free paths, Notation:

[0050] (12);

[0051] Among them, in the allocation scheme The CCP For intelligent connected vehicles, therefore, the set of collision-free paths There is Okay, representing There are several collision-free paths, each containing several path points at different times; For each path point in the set of collision-free paths, denoted as That is, the first The set of collision-free paths for vehicles at time steps The coordinates of the location in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. For time steps, The maximum number of steps;

[0052] definition Indicating intelligent connected vehicles exist Collision-free paths within a time step, using Indicates all Intelligent connected vehicles in time step The set of path points; the set of collision-free paths Matches: Target location of intelligent connected vehicles The allocation matrix should correspond to the allocation scheme. The elements correspond to each other, and at each time step set of path points Neither should contain micro-level conflict relationships;

[0053] First, utilize path planning The algorithm is based on the current allocation matrix. Perform path planning to obtain an initial set of collision-free paths. Check the current set of collision-free paths. Path points with micro-conflicts Add path points with micro-conflicts to the corresponding conflict point set. In the middle, the conflict points will be set As a constraint and utilizing path planning The algorithm executes the next round of path planning, repeatedly adding micro-conflicting waypoints and performing path planning again, until a set of collision-free paths is reached. When there are no more path points with microscopic conflicts in the path set, that is, when there are no more path points with microscopic conflicts in the path set, the path set is considered a collision-free path set. The process of eliminating microscopic conflicts has been completed, the iteration is now finished, and a collision-free path for each intelligent connected vehicle has been obtained.

[0054] A3.3: Current Iteration Index Add 1;

[0055] A3.4: The allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, is taken as the suboptimal allocation scheme for the current iteration. The lane-changing physical cost of the suboptimal allocation scheme is compared with the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme. If the lane-changing physical cost of the suboptimal allocation scheme is greater than the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme, then the collision-free path set corresponding to the optimal allocation scheme is taken as the optimal solution and A3.9 is executed; otherwise, A3.5 is executed.

[0056] A3.5: Perform path planning on the suboptimal allocation scheme of the current iteration to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs;

[0057] A3.6: Current Iteration Index Add 1;

[0058] A3.7: Take the allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, as the suboptimal allocation scheme for the current iteration, perform path planning to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs, and return to A3.6, until all allocation schemes are traversed and A3.8 is executed.

[0059] A3.8: The set of collision-free paths that minimize the actual physical cost of lane changing is taken as the optimal solution;

[0060] A3.9: Use the optimal solution as the final set of collision-free paths obtained from the solution.

[0061] Furthermore, the method of controlling intelligent connected vehicles to change lanes in the lane-changing zone and enter the car-following zone according to the collision-free path set, while satisfying car-following safety and physical kinematic constraints, plans the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, specifically including the following steps:

[0062] C1: Intelligent connected vehicles enter the following zone; determine the entry time of each intelligent connected vehicle into the following zone. The initial speed when entering the car-following zone and the assigned target entry time into the intersection ;

[0063] C2: Establish a decentralized energy-optimal control model for intelligent connected vehicles;

[0064] For intelligent connected vehicles The decentralized energy optimal control model is established as follows:

[0065] (13);

[0066] in, This is the input for vehicle control, namely acceleration;

[0067] C3: Solve the decentralized energy-optimal control model to obtain the optimal acceleration trajectory at the micro level for each intelligent connected vehicle;

[0068] To solve the decentralized energy-optimal control model, a Hamiltonian function is constructed using the Pontryagin maximum principle. for:

[0069] (14);

[0070] in, For the position, speed, and acceleration of intelligent connected vehicles, and For costate variables, For the Lagrange multipliers corresponding to the constraints, For acceleration constraints, let represent the maximum and minimum acceleration values, respectively; For speed constraints, let represent the maximum and minimum speed values, respectively;

[0071] According to the Euler-Lagrange equations, we can obtain:

[0072] (15);

[0073] (16);

[0074] in, for The derivative, for The derivative, For the speed of intelligent connected vehicles;

[0075] Let be a constant, and let it be... Integrating it yields ,in Let be the integration constant, and then according to the necessary condition for optimality:

[0076] (17);

[0077] The optimal control law under conditions without velocity and acceleration constraints is derived to be a linear function of time.

[0078] (18);

[0079] in, The optimal acceleration;

[0080] Substituting this optimal acceleration into the vehicle dynamics equations and performing a quadratic integration yields the energy-optimal velocity at the microscopic level. With position Trajectory Programming Equation:

[0081] (19);

[0082] (20);

[0083] in, It is the integration constant;

[0084] The microscopic kinematic boundary conditions are set as follows: the initial velocity entering the catenary region. Initial position for entering the following zone Terminal location The boundary conditions for the costate variables are set as follows: , and Indicates the time when the assigned target enters the intersection. Costate variables under;

[0085] A system of linear equations is constructed based on the boundary conditions of micro-kinematics and co-state variables, and the undetermined coefficients are solved in real time. :

[0086] (twenty one);

[0087] in, This is a transpose.

[0088] Furthermore, the process of planning the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, driving the intelligent connected vehicle to the stop line, and using the graph relationship method to achieve conflict-free passage in the scheduling area specifically includes the following steps:

[0089] D1: Model all intelligent connected vehicles and construct a conflict directed graph, which includes several nodes, one node corresponds to one intelligent connected vehicle, and the edges represent the conflict relationships between intelligent connected vehicles.

[0090] Define a conflict-directed graph The set of nodes in a conflicting directed graph is defined as follows: The set of edges of a conflicting directed graph It consists of the union of the set of one-way edges and the set of two-way edges:

[0091] (twenty two);

[0092] in, It is a set of one-way edges. It is a set of bidirectional edges;

[0093] The one-way edges in the set of one-way edges represent car-following constraints, occurring between two connected vehicles entering the intersection from the same lane; if there exists a car-following constraint in the conflict directed graph from node... Pointing to node A one-way edge means that in physical space, the node Absolutely cannot compare to nodes First reach the conflict point at the intersection;

[0094] (twenty three);

[0095] in, The set of traffic splitting conflicts is the set of intelligent connected vehicles that have a car-following constraint with the intelligent connected vehicle. Represents an edge;

[0096] The bidirectional edges in the bidirectional edge set represent cross-conflict constraints with a travel order game space. When there is a spatial intersection of the predetermined travel trajectories of intelligent connected vehicles in different directions or different lanes, if there is a bidirectional edge between two nodes, it means that the arrival order of the two intelligent connected vehicles can be interchanged, but they absolutely cannot occupy the conflict point at the same time.

[0097] (twenty four);

[0098] in, The cross-conflict set is the set of intelligent connected vehicles that have cross-conflicts with the intelligent connected vehicle.

[0099] D2: The complement of the conflicting directed graph is taken as the coexisting undirected graph. By solving the maximum matching problem in the coexisting undirected graph, the vehicle pairing scheme with the highest spatiotemporal resource utilization at the intersection is obtained, so as to realize the conflict-free passage in the scheduling area. The coexisting undirected graph includes several vertices and edges. Each vertex represents an intelligent connected vehicle, and each edge represents that there is no conflict relationship between two vertices.

[0100] On the other hand, the present invention also provides a multi-lane unsignalized intersection connected vehicle scheduling system that considers lane changes, for implementing a multi-lane unsignalized intersection connected vehicle scheduling method that considers lane changes, including:

[0101] The area division module is used to determine the upstream road segment area of ​​an unsignalized intersection and construct the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. It spatially divides the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions, namely the lane changing area and the following area.

[0102] The mapping module is used to divide the road into discrete grid nodes in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, and to map the actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, so as to obtain the initial relative coordinates of each intelligent connected vehicle and determine the target lane of each intelligent connected vehicle.

[0103] The lane change zone allocation scheme and collision-free path solution module is used to model the target lane allocation as a 0-1 integer allocation planning problem. By minimizing the physical cost of lane changing for the entire fleet, the 0-1 integer allocation planning problem is solved iteratively to obtain the allocation scheme and its corresponding physical cost for each iteration. The collision-free path set of the allocation scheme is solved in ascending order of physical cost of lane changing, and finally the collision-free path set with the minimum actual physical cost of lane changing is obtained.

[0104] The optimal acceleration trajectory solution module for the car-following zone is used to control intelligent connected vehicles to change lanes in the lane-changing zone and enter the car-following zone according to the set of collision-free paths. Under the premise of satisfying car-following safety and physical kinematic constraints, it plans the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times.

[0105] The dispatch area passage module is used to plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, so that the intelligent connected vehicle can drive to the stop line and achieve conflict-free passage in the dispatch area using the graph relationship method.

[0106] Thirdly, this application proposes an electronic device comprising: one or more processors, and a memory for storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the aforementioned multi-lane unsignalized intersection connected vehicle scheduling method considering lane changes.

[0107] Fourthly, this application proposes a computer-readable storage medium storing executable instructions that, when executed, cause a processor to perform the described multi-lane unsignaled intersection vehicle scheduling method considering lane changes.

[0108] Fifthly, this application proposes a computer program product, including a computer program or instructions that, when executed by a processor, implement the aforementioned method for scheduling connected vehicles at multi-lane unsignalized intersections that takes lane changes into account.

[0109] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0110] Dimensional reduction and decoupling improve efficiency: In the spatial dimension, this invention utilizes the lane-changing zone to absorb the initial randomness of traffic flow; in the temporal dimension, it utilizes the following zone to extract the maximum parallel traffic potential of the intersection, which greatly optimizes the overall evacuation efficiency of the intersection.

[0111] Avoid congestion and deadlock: By actively coordinating lane changes at a remote location, the extreme imbalance in traffic distribution between lanes is eliminated, effectively avoiding long queues in single lanes caused by random vehicle distribution.

[0112] Energy saving, emission reduction and smooth control: At the micro level, based on the optimal control model, a smooth trajectory with optimal energy is planned for the vehicle, which minimizes the energy loss caused by frequent acceleration and deceleration. Attached Figure Description

[0113] Figure 1 This is a flowchart of a multi-lane unsignalized intersection vehicle scheduling method considering lane changes, as described in an embodiment of the present invention.

[0114] Figure 2 This is a diagram showing the intersection area division in an embodiment of the present invention;

[0115] Figure 3 This is a schematic diagram of the relative coordinate system of the upstream road segment area of ​​an unsignalized intersection in an embodiment of the present invention;

[0116] Figure 4 This is a flowchart of the path planning iterative optimization process in an embodiment of the present invention;

[0117] Figure 5 This is a directed graph showing conflict relationships in an embodiment of the present invention;

[0118] Figure 6 This is a conflict relationship diagram of the scheduling area in an embodiment of the present invention;

[0119] Figure 7 This is an undirected graph showing the coexistence relationships in an embodiment of the present invention. Detailed Implementation

[0120] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0121] This invention proposes a method for scheduling connected vehicles at multi-lane, unsignaled intersections that considers lane changes. It changes the idealized assumption of fixed lanes being unchangeable in a purely connected environment, extending the scope of intersection scheduling to upstream road segments. By establishing a two-stage collaborative control framework of lane-changing zone and car-following zone, the first stage utilizes 0-1 integer programming and conflict search algorithms to actively allocate target lanes and plan conflict-free trajectories for intelligent connected vehicles (CAVs), achieving dynamic equilibrium of multi-lane spatial resources. In the second stage, combined with the proposed micro-energy optimal control strategy, the optimal travel time is transformed into a smooth and coherent physical motion trajectory for the underlying vehicles, such as... Figure 1 As shown, the method specifically includes the following steps:

[0122] S1: Determine the upstream road segment area of ​​the unsignalized intersection and construct the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. Spatially divide the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions, namely the lane-changing area and the following area.

[0123] Specifically, such as Figure 2 As shown, the central area within the stop line of the unsignalized intersection is first designated as the dispatch zone, and the length of the upstream road segment of the unsignalized intersection is set to... The distance from the parking line outside the dispatch area is The area upstream of the unsignalized intersection is strictly divided into a lane-changing zone and a following zone in order to manage the lane changing and intersection passage of intelligent connected vehicles. The lengths of these zones are as follows: and And satisfy The lane-changing zone is located on the outermost edge of the upstream section of the unsignalized intersection, where connected vehicles can change lanes. The following zone is the core area adjacent to the dispatch zone, where connected vehicles are prohibited from changing lanes and can only follow the connected vehicle in front in the designated target lane.

[0124] To transform continuous two-dimensional planar lane-changing trajectories into computable discrete mathematical problems, such as... Figure 3 As shown, the origin is a point located on the innermost lane, on the boundary line between the lane-changing area and the following lane, and the direction away from the intersection is defined as... In the axial direction, from the innermost lane to the outermost lane. In the axial direction, a relative coordinate system is constructed for the upstream road segment area of ​​the unsignalized intersection;

[0125] S2: In the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, the road is divided into discrete grid nodes. The actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area are mapped to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection to obtain the initial relative coordinates of each intelligent connected vehicle. This provides a unified discrete calculation framework for subsequent algorithms and determines the target lane of each intelligent connected vehicle.

[0126] Specifically:

[0127] When a batch of intelligent connected vehicles arrives at the starting point of the lane-changing zone, their physical locations are first mapped to the nearest discrete grid node, and the intelligent connected vehicles are then configured. The actual two-dimensional physical coordinates are Its mapped coordinates in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection are: It is obtained by solving the following problem of minimizing Euclidean distance:

[0128] (1);

[0129] in, The set of natural numbers, A relative coordinate system representing the upstream road segment area of ​​an unsignalized intersection; through the above mapping, each intelligent connected vehicle obtains a unique initial relative coordinate.

[0130] At the same time, at the boundary of the lane-changing zone, there are also a number of target location They are to be allocated to various vehicles. The number of vehicles;

[0131] S3: Based on the target lane of each intelligent connected vehicle, the target lane allocation is modeled as a 0-1 integer allocation planning problem. The 0-1 integer allocation planning problem is solved iteratively by minimizing the physical cost of lane changing for the entire fleet. The allocation scheme and its corresponding physical cost of lane changing are obtained in each iteration. The collision-free path of the allocation scheme is solved in order of increasing physical cost of lane changing, and finally the collision-free path with the minimum physical cost of lane changing is obtained.

[0132] S3.1: Construct the lane-changing physical cost matrix, preference matrix, and allocation matrix;

[0133] Before assigning a target location to an intelligent connected vehicle, the lane-changing physical cost of that assignment should be defined. In this invention, the distance from the initial relative coordinates to the target location is defined as the lane-changing physical cost of a single intelligent connected vehicle, and Euclidean distance is used as the evaluation metric. The following lane-changing physical cost matrix can then be obtained:

[0134] (2);

[0135] (3);

[0136] in, For the physical cost matrix of lane changing, For the lane-changing physical cost matrix, the first... Line number The elements of the column represent intelligent connected vehicles. Assigned to target location The cost of allocation The target location number. Represents the set of positive integers;

[0137] In this embodiment, according to Figure 3 In the medium case, the calculated lane-switching physical cost matrix is:

[0138] (4);

[0139] Based on the target lane of each intelligent connected vehicle, the preferences of the intelligent connected vehicles are defined by a preference matrix, expressed as follows:

[0140] (5);

[0141] (6);

[0142] in, For the preference matrix, The first in the preference matrix Line number The column elements represent intelligent connected vehicles taking lane preference into account. Can it be assigned to the target location? ;

[0143] In this embodiment, according to Figure 3 In the middle case, the calculated preference matrix is:

[0144] (7);

[0145] Finally, define the Boolean decision variables, i.e., the assignment matrix, as follows:

[0146] (8);

[0147] (9);

[0148] in, For the allocation matrix, For the allocation matrix, the first Line number The elements of the column represent intelligent connected vehicles. Whether successfully assigned to the target location If so Otherwise, it is 0;

[0149] Since each intelligent connected vehicle can only be assigned one target location, There are necessary constraints: each column and each row can have only one element that is 1.

[0150] S3.2: Based on the lane-switching physical cost matrix, preference matrix, and allocation matrix, construct a 0-1 integer programming allocation problem and solve it iteratively to obtain the allocation scheme and its corresponding lane-switching physical cost for each iteration;

[0151] The 0-1 integer programming allocation problem is as follows:

[0152] (10);

[0153] Finally, by solving the 0-1 integer programming assignment problem, the assignment matrix is ​​obtained. The Hungarian algorithm is often used to solve this type of allocation problem, so this invention uses the algorithm to solve the 0-1 integer programming allocation problem.

[0154] S3.3: Based on the allocation scheme of each iteration and its corresponding lane-changing physical cost, solve for the set of collision-free paths with the minimum actual lane-changing physical cost;

[0155] like Figure 4 As shown, the specific steps include:

[0156] S3.3.1: Set the initial iteration index =1;

[0157] S3.3.2: Obtain the allocation scheme with the minimum lane-changing physical cost as the optimal allocation scheme, perform path planning, and solve for its collision-free path set and the actual lane-changing physical cost. The collision-free path set includes the collision-free path of each intelligent connected vehicle.

[0158] In this embodiment, the allocation matrix corresponding to the optimal allocation scheme obtained by the Hungarian algorithm is... for:

[0159] (11);

[0160] This means that intelligent connected vehicle 1 is assigned to target location 3, intelligent connected vehicle 2 is assigned to target location 4, and so on, with a corresponding lane-changing physical cost of 18.73.

[0161] The allocation matrix corresponding to the optimal allocation scheme is obtained. It only guarantees that the physical cost of lane changing between the initial relative coordinates and the target position is minimized, but does not consider whether dynamic collisions will occur when the intelligent connected vehicle moves laterally to the target position. Therefore, collision-free path planning must be performed to obtain a collision-free path.

[0162] In the relative coordinate system of the upstream section of an unsignalized intersection, intelligent connected vehicles are restricted to moving only in orthogonal directions (diagonal movement is prohibited to prevent large overload acceleration and deceleration). Intelligent connected vehicles may encounter the following three micro-conflicts during concurrent lane changes:

[0163] (1) Node conflict: Intelligent connected vehicles attempt to occupy the same discrete grid node as their target location at the same time step;

[0164] (2) Edge conflict: Two intelligent connected vehicles exchange paths in opposite directions between adjacent discrete grid nodes at the same time step (i.e., trajectory crossing).

[0165] (3) Intermediate conflict: Considering the physical size of intelligent connected vehicles and the time difference of lane changing, an intelligent connected vehicle forcibly cuts in before another intelligent connected vehicle has completely left the discrete grid node;

[0166] The collision-free path planning process is as follows:

[0167] Similar to single-agent path planning, the main task of intelligent connected vehicles in the lane-changing zone is to successfully change lanes to the generated target location, i.e., to generate a collision-free path according to the target allocation scheme. Therefore, this invention uses conflict-based search to solve the multi-vehicle path planning problem. The current allocation scheme for the next iteration ,definition For the corresponding set of collision-free paths, Notation:

[0168] (12);

[0169] Among them, in the allocation scheme The CCP For intelligent connected vehicles, therefore, the set of collision-free paths There is Okay, representing There are several collision-free paths, each containing several path points at different times; For each path point in the set of collision-free paths, denoted as That is, the first The set of collision-free paths for vehicles at time steps The coordinates of the location in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. For time steps; considering the limited length of the lane change zone, define The maximum number of steps;

[0170] definition Indicating intelligent connected vehicles exist Collision-free paths within a time step, using Indicates all Intelligent connected vehicles in time step The set of path points; the set of collision-free paths It should meet the following criteria: the target location of the intelligent connected vehicle. The allocation matrix should correspond to the allocation scheme. The elements correspond to each other, and at each time step set of path points Neither should contain micro-level conflict relationships;

[0171] First, utilize path planning The algorithm is based on the current allocation matrix. Perform path planning to obtain an initial set of collision-free paths. Check the current set of collision-free paths. Path points with micro-conflicts Add path points with micro-conflicts to the corresponding conflict point set. In the middle, the conflict points will be set As a constraint and utilizing path planning The algorithm executes the next round of path planning, repeatedly adding micro-conflicting waypoints and performing path planning again, until a set of collision-free paths is reached. When there are no more path points with microscopic conflicts in the path set, that is, when there are no more path points with microscopic conflicts in the path set, the path set is considered a collision-free path set. The iteration has ended, achieving zero micro-conflicts and yielding a collision-free path for each connected vehicle. However, since connected vehicles experiencing micro-conflicts are equal in status, there are still path points with micro-conflicts. It can be added to the set of conflict points of either side. middle;

[0172] However, the allocation matrix corresponding to the theoretically optimal allocation scheme obtained by relying solely on a single algorithm is... The theoretical physical cost of lane changing is The actual physical cost of lane changing generated by path planning is The set of collision-free paths This is only a locally optimal solution because another suboptimal allocation scheme, although theoretically slightly farther, may have a lower actual physical cost of changing routes due to fewer conflicts and no need for detours. Therefore, path planning should continue.

[0173] S3.3.3: Current Iteration Index Add 1;

[0174] S3.3.4: The allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, is taken as the suboptimal allocation scheme for the current iteration. The lane-changing physical cost of the suboptimal allocation scheme is compared with the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme. If the lane-changing physical cost of the suboptimal allocation scheme is greater than the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme, then the collision-free path set corresponding to the optimal allocation scheme is taken as the optimal solution and S3.3.9 is executed; otherwise, S3.3.5 is executed.

[0175] In this embodiment, the Hungarian algorithm is used to generate the allocation matrix corresponding to the suboptimal allocation scheme. The physical cost of its lane change is in addition to The lowest among all allocation matrices except for, due to The allocation matrix is ​​not the one corresponding to the (initial) optimal allocation scheme, therefore the cost satisfies... ( > 2);

[0176] Therefore, it is obvious that if Then the set of collision-free paths It is the globally optimal solution because it is based on the set of collision-free paths. The actual physical cost of lane changing generated Must be greater than Furthermore, the actual physical cost of lane changing for all remaining allocation schemes will be greater;

[0177] S3.3.5: Perform path planning on the suboptimal allocation scheme of the current iteration to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs;

[0178] S3.3.6: Current Iteration Index Add 1;

[0179] S3.3.7: Take the allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, as the suboptimal allocation scheme for the current iteration, perform path planning to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs, and return to S3.3.6, until all allocation schemes are traversed and S3.3.8 is executed.

[0180] S3.3.8: The set of collision-free paths that minimize the actual physical cost of lane changing is taken as the optimal solution;

[0181] S3.3.9: Use the optimal solution as the final set of collision-free paths obtained from the solution;

[0182] S4: Control the intelligent connected vehicles to change lanes in the lane-changing area and enter the car-following area according to the collision-free path set. Under the premise of satisfying car-following safety and physical kinematic constraints, plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times.

[0183] S4.1: After upstream lane-changing allocation and macro-level right-of-way scheduling, intelligent connected vehicles enter the car-following zone. At this point, the entry time of each intelligent connected vehicle into the car-following zone is determined. The initial speed when entering the car-following zone and the assigned target entry time into the intersection ;

[0184] S4.2: Establish a decentralized energy-optimal control model for intelligent connected vehicles;

[0185] Since internal combustion engines or drive motors operate at their highest efficiency under steady-state conditions, frequent acceleration and deceleration (transient operations) lead to significant energy loss. Therefore, in order to minimize fuel consumption and emissions at the micro-control level, the objective function of micro-cooperative control is set as the vehicle control input within the time interval. Norms, for intelligent connected vehicles The decentralized energy optimal control model is established as follows:

[0186] (13);

[0187] in, This is the input for vehicle control, namely acceleration;

[0188] S4.3: Solve the decentralized energy-optimal control model to obtain the optimal acceleration trajectory at the micro level for each intelligent connected vehicle;

[0189] To solve the aforementioned continuous-time decentralized energy-optimal control model, a Hamiltonian function is constructed using the Pontryagin maximum principle. State and control constraints are introduced as penalty terms, resulting in the constructed Hamiltonian function. for:

[0190] (14);

[0191] in, For the position, speed, and acceleration of intelligent connected vehicles, and For costate variables, For the Lagrange multipliers corresponding to the constraints, For acceleration constraints, let represent the maximum and minimum acceleration values, respectively; For speed constraints, let represent the maximum and minimum speed values, respectively. Under ideal and congestion-free smooth scheduling conditions, the speed and acceleration of intelligent connected vehicles within the following zone do not reach physical limits; that is, neither acceleration nor speed constraints are activated. ;

[0192] According to the Euler-Lagrange equations, we can obtain:

[0193] (15);

[0194] (16);

[0195] in, for The derivative, for The derivative, For the speed of intelligent connected vehicles;

[0196] From the above partial derivative equations, we can see that Let be a constant, and let it be... Integrating it yields ,in Let be the integration constant, and then according to the necessary condition for optimality:

[0197] (17);

[0198] It can be deduced that the optimal control law in the state without velocity and acceleration constraints is a linear function of time:

[0199] (18);

[0200] in, The optimal acceleration;

[0201] Substituting this optimal acceleration into the vehicle dynamics equations and performing a quadratic integration yields the energy-optimal velocity at the microscopic level. With position Trajectory Programming Equation:

[0202] (19);

[0203] (20);

[0204] in, It is the integration constant;

[0205] Since the microscopic kinematic boundary conditions are known: the initial velocity entering the cathodic region Initial position for entering the following zone Terminal location Furthermore, to ensure that intelligent connected vehicles smoothly enter intersections, costate variable boundary conditions are typically set. , and Indicates the time when the assigned target enters the intersection. Costate variables under;

[0206] Substituting the microscopic kinematic boundary conditions and co-state variable boundary conditions into the above equations allows for the construction of a system of linear equations and the real-time solution of the undetermined coefficients. :

[0207] (twenty one);

[0208] in, For transpose;

[0209] S5: Based on the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, the intelligent connected vehicle travels to the parking line and uses the graph relationship method to achieve conflict-free passage in the scheduling area.

[0210] S5.1: Model all intelligent connected vehicles and construct a conflict directed graph to describe the conflict topology of intelligent connected vehicles in the scheduling area. This graph includes several nodes, with each node corresponding to an intelligent connected vehicle, and edges representing the conflict relationships between intelligent connected vehicles.

[0211] Define a conflict-directed graph The set of nodes in a conflicting directed graph is defined as follows: The set of edges of a conflicting directed graph It consists of the union of the set of one-way edges and the set of two-way edges, such as Figure 5 ,Right now:

[0212] (twenty two);

[0213] in, It is a set of one-way edges. It is a set of bidirectional edges;

[0214] The one-way edges in the set of one-way edges represent car-following constraints, occurring between two connected vehicles entering the intersection from the same lane; if there exists a car-following constraint in the conflict directed graph from node... Pointing to node A one-way edge means that in physical space, the node Absolutely cannot compare to nodes First reach the conflict point at the intersection;

[0215] (twenty three);

[0216] in, The set of traffic splitting conflicts is the set of intelligent connected vehicles that have a car-following constraint with the intelligent connected vehicle. Represents an edge;

[0217] The bidirectional edges in the set represent cross-conflict constraints in a game space with a travel order. When the predetermined travel trajectories of intelligent connected vehicles in different directions or lanes intersect, if a bidirectional edge exists between two nodes, it means that the arrival order of these two intelligent connected vehicles can be interchanged (i.e., ...). You can go first. (They can go first), but they absolutely cannot occupy the conflict point at the same time;

[0218] (twenty four);

[0219] in, The cross-conflict set is the set of intelligent connected vehicles that have cross-conflicts with the intelligent connected vehicle.

[0220] To visually illustrate the construction process of the conflict-directed graph, consider a typical intersection scenario involving 6 intelligent connected vehicles. See... Figure 6 Assume that intelligent connected vehicle 1 and intelligent connected vehicle 2 are located in the straight and left-turn lanes of the west entrance, respectively; intelligent connected vehicle 3 and intelligent connected vehicle 4 are located in the straight lanes of the north and east entrances, respectively; and intelligent connected vehicle 5 and intelligent connected vehicle 6 are both located in the straight lanes of the south entrance (vehicle 5 in front, vehicle 6 behind).

[0221] Based on the above rules, the conflict set of each intelligent connected vehicle can be obtained as follows: , ; , ; , ; , ; , ; , ;

[0222] S5.2: The complement of the conflicting directed graph is taken as the coexisting undirected graph. By solving the maximum matching problem in the coexisting undirected graph, the vehicle pairing scheme with the highest spatiotemporal resource utilization at the intersection is obtained, so as to realize the conflict-free passage in the scheduling area. The coexisting undirected graph includes several vertices and edges. Each vertex represents an intelligent connected vehicle, and each edge represents that there is no conflict relationship between two vertices.

[0223] In scheduling algorithms, minimizing the total evacuation time of all vehicles is equivalent to minimizing the total depth of the vehicle spanning tree. Since the total number of vehicles in the control area is fixed, reducing the depth of the spanning tree means maximizing its width. In other words, we need to find the maximum number of vehicle combinations that can simultaneously pass through the intersection without conflict.

[0224] In graph theory, this objective is equivalent to solving coexisting undirected graphs. The maximum matching problem in the equation. It is a coexisting undirected graph A subset of the middle edge set, see Figure 7 In this subset, no two edges share the same vertex; the physical meaning here is that matching... Each edge in the diagram represents two non-conflicting intelligent connected vehicles that can pass simultaneously, and each vehicle is only scheduled once in the same batch;

[0225] If for any other matching in the graph All satisfied Then it is called For the maximum matching, denoted as This corresponds to the vehicle pairing scheme that maximizes the utilization of time and space resources at the intersection.

[0226] In summary, this invention breaks through the limitations of the ideal assumption of prohibiting lane changes in traditional intersection scheduling, and proposes a two-stage collaborative control architecture of upstream macro-allocation and downstream micro-scheduling. Spatially, by constructing a relative coordinate system in the lane-changing zone and combining iterative solutions of target allocation and conflict-free path planning, initial equilibrium of multi-lane traffic flow is achieved, effectively avoiding local queuing and deadlocks caused by uneven vehicle distribution. Temporally, in the car-following zone, graph theory methods are used to transform complex micro-physical constraints into mathematical topological structures, accurately calculating the optimal traffic sequence at the intersection. Furthermore, this invention constructs an energy-optimal control model based on the maximum principle, planning smooth micro-motion trajectories for vehicles. This method not only greatly improves the spatial resource utilization and overall evacuation efficiency of intersections, but also significantly reduces frequent acceleration and deceleration operations of vehicles while ensuring traffic safety, effectively reducing the fuel consumption and emissions of the entire system.

[0227] Example 2:

[0228] A networked vehicle dispatching system for multi-lane unsignalized intersections that considers lane changes, used to implement a networked vehicle dispatching method for multi-lane unsignalized intersections that considers lane changes, including:

[0229] The area division module is used to determine the upstream road segment area of ​​an unsignalized intersection and construct the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. It spatially divides the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions, namely the lane changing area and the following area.

[0230] The mapping module is used to divide the road into discrete grid nodes in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, and to map the actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, so as to obtain the initial relative coordinates of each intelligent connected vehicle and determine the target lane of each intelligent connected vehicle.

[0231] The lane change zone allocation scheme and collision-free path solution module is used to model the target lane allocation as a 0-1 integer allocation planning problem. By minimizing the physical cost of lane changing for the entire fleet, the 0-1 integer allocation planning problem is solved iteratively to obtain the allocation scheme and its corresponding physical cost for each iteration. The collision-free path set of the allocation scheme is solved in ascending order of physical cost of lane changing, and finally the collision-free path set with the minimum actual physical cost of lane changing is obtained.

[0232] The optimal acceleration trajectory solution module for the car-following zone is used to control intelligent connected vehicles to change lanes in the lane-changing zone and enter the car-following zone according to the set of collision-free paths. Under the premise of satisfying car-following safety and physical kinematic constraints, it plans the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times.

[0233] The dispatch area passage module is used to plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, so that the intelligent connected vehicle can drive to the stop line and achieve conflict-free passage in the dispatch area using the graph relationship method.

[0234] Example 3:

[0235] This embodiment proposes an electronic device, including: one or more processors, and a memory, the memory being used to store instructions, which, when executed by the one or more processors, cause the one or more processors to execute the multi-lane unsignalized intersection connected vehicle scheduling method considering lane changes.

[0236] The electronic device may be a mobile phone, computer, or tablet computer, etc., and includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the multi-lane unsignalized intersection connected vehicle scheduling method considering lane changes as described in the embodiments. It is understood that the electronic device may also include input / output (I / O) interfaces and communication components.

[0237] The processor is used to execute all or part of the steps in the multi-lane unsignalized intersection connected vehicle scheduling method considering lane changes as described in the above embodiments. The memory is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.

[0238] The processor can be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic components, and is used to execute the multi-lane unsignalized intersection vehicle scheduling method considering lane changes described in the above embodiments.

[0239] Example 4:

[0240] This embodiment proposes a computer-readable storage medium that stores executable instructions. When these instructions are executed, if they are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0241] The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the multi-lane unsignalized intersection vehicle scheduling method considering lane changes described in the various embodiments of this application.

[0242] The aforementioned storage media include: flash memory, hard disk, multimedia card, card-type memory (e.g., SD (Secure Digital Memory Card) or DX (Memory Data Register, MDR) memory), random access memory (RAM), static random-access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, disk, optical disk, server, APP (Application) application store, and other media capable of storing program verification codes. These media store computer programs, which, when executed by a processor, can implement the various steps of the aforementioned multi-lane unsignalized intersection network vehicle scheduling method considering lane changes.

[0243] Example 5:

[0244] This embodiment proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the aforementioned multi-lane unsignalized intersection connected vehicle scheduling method that considers lane changes.

[0245] Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a computer program product.

[0246] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0247] The scope of protection of this application is not limited to the embodiments described above. Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from the scope and spirit of this disclosure. If such modifications and variations fall within the scope of this disclosure and its equivalents, then the intent of this disclosure also includes these modifications and variations.

[0248] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for scheduling connected vehicles at a multi-lane, unsignaled intersection considering lane changes, characterized in that, Includes the following steps: Determine the upstream road segment area of ​​the unsignalized intersection and construct a relative coordinate system for the upstream road segment area of ​​the unsignalized intersection. Spatially divide the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions: the lane-changing area and the following area. In the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, the road is divided into discrete grid nodes. The actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area are mapped to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection to obtain the initial relative coordinates of each intelligent connected vehicle and determine the target lane of each intelligent connected vehicle. The target lane assignment is modeled as a 0-1 integer assignment planning problem. The 0-1 integer assignment planning problem is solved iteratively by minimizing the physical cost of lane changing for the entire fleet. The assignment scheme and its corresponding physical cost for each iteration are obtained. The set of collision-free paths for the assignment schemes is solved in ascending order of physical cost of lane changing. Finally, the set of collision-free paths with the minimum physical cost of lane changing is obtained, including the collision-free paths of each intelligent connected vehicle. According to the collision-free path set, the intelligent connected vehicles change lanes in the lane change zone and enter the following zone, and plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times. Based on the optimal acceleration trajectory planned at the micro level for each intelligent connected vehicle, the intelligent connected vehicle travels to the parking line and uses the graph relationship method to achieve conflict-free passage in the scheduling area.

2. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 1, characterized in that, The process involves determining the upstream road segment area of ​​the unsignalized intersection and constructing a relative coordinate system for the upstream road segment area. This divides the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions: a lane-changing area and a following-car area. Specifically: First, the central area within the stop line of the unsignalized intersection is designated as the dispatch zone, and the length of the upstream road segment of the unsignalized intersection is set as follows: The distance from the parking line outside the dispatch area is The area upstream of the unsignalized intersection is spatially divided into a lane-changing zone and a following zone, with lengths of [missing information]. and And satisfy The lane-changing zone is located on the outermost edge of the upstream section of the unsignalized intersection. Intelligent connected vehicles change lanes in the lane-changing zone and enter the target lane. The following zone is the area adjacent to the dispatch zone. Intelligent connected vehicles are prohibited from changing lanes and must follow the intelligent connected vehicle in front in the designated target lane. Take the point located on the innermost lane, on the boundary line between the lane-changing area and the following lane, as the origin, and take the direction away from the intersection as... In the axial direction, from the innermost lane to the outermost lane. Along the axis, a relative coordinate system is constructed for the upstream road segment area of ​​the unsignalized intersection.

3. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 1, characterized in that, In the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, the road is divided into discrete grid nodes. The actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area are mapped to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, thus obtaining the initial relative coordinates of each intelligent connected vehicle. At the same time, the target lane of each intelligent connected vehicle is determined, specifically as follows: Setting up intelligent connected vehicles The actual two-dimensional physical coordinates are The mapped coordinates of the upstream road segment of an unsignalized intersection in the relative coordinate system are obtained by solving the following problem of minimizing Euclidean distance. : （1）； in, The set of natural numbers, A relative coordinate system representing the upstream road segment area of ​​an unsignalized intersection; At the same time, at the boundary of the lane-changing zone, there are also a number of target location They are to be allocated to various vehicles. The number of vehicles.

4. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 1, characterized in that, The target lane assignment is modeled as a 0-1 integer assignment planning problem. The problem is iteratively solved by minimizing the physical cost of lane changes for the entire platoon, yielding the assignment scheme and its corresponding physical cost for each iteration. The set of collision-free paths for each assignment scheme is then calculated sequentially according to the physical cost of lane changes, from smallest to largest. Finally, the set of collision-free paths with the minimum actual physical cost of lane changes is obtained. This process includes the following steps: A1: Construct the lane-changing physical cost matrix, preference matrix, and allocation matrix; The lane-changing physical cost matrix is ​​as follows: （2）； （3）； in, For the physical cost matrix of lane changing, For the lane-changing physical cost matrix, the first... Line number The elements of the column represent intelligent connected vehicles. Assigned to target location The cost of allocation The target location number. Represents the set of positive integers; Based on the target lane of each intelligent connected vehicle, the preferences of the intelligent connected vehicles are defined by a preference matrix, expressed as follows: （5）； （6）； in, For the preference matrix, The first in the preference matrix Line number The column elements represent intelligent connected vehicles taking lane preference into account. Can it be assigned to the target location? ; The allocation matrix is ​​represented as follows: （8）； （9）； in, For the allocation matrix, For the allocation matrix, the first Line number The elements of the column represent intelligent connected vehicles. Whether successfully assigned to the target location ; Allocation matrix There is a necessary constraint: each column and each row can have only one element that is 1; A2: Based on the lane-switching physical cost matrix, preference matrix, and allocation matrix, construct a 0-1 integer programming allocation problem and solve it iteratively to obtain the allocation scheme and its corresponding lane-switching physical cost for each iteration; The 0-1 integer programming allocation problem is as follows: （10）； Solve the 0-1 integer programming assignment problem to obtain the assignment matrix. ; A3: Based on the allocation scheme of each iteration and its corresponding lane-changing physical cost, solve for the set of collision-free paths with the minimum actual lane-changing physical cost.

5. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 4, characterized in that, The A3 specifically includes the following steps: A3.1: Setting the initial iteration index =1; A3.2: Obtain the allocation scheme with the minimum lane-changing physical cost as the optimal allocation scheme, perform path planning, and solve for its collision-free path set and actual lane-changing physical cost; In the relative coordinate system of the upstream section of an unsignalized intersection, intelligent connected vehicles are restricted to moving only in orthogonal directions. During concurrent lane changes, intelligent connected vehicles face the following three micro-conflicts: (1) Node conflict: Intelligent connected vehicles attempt to occupy the same discrete grid node as their target location at the same time step; (2) Edge conflict: Two intelligent connected vehicles exchange with each other in the same time step between adjacent discrete grid nodes; (3) Intermediate conflict: Considering the physical size of intelligent connected vehicles and the time difference of lane changing, an intelligent connected vehicle forcibly cuts in before another intelligent connected vehicle has completely left the discrete grid node; The collision-free path planning process is as follows: For the The current allocation scheme for the next iteration ,definition For the corresponding set of collision-free paths, Notation: （12）； Among them, in the allocation scheme The CCP For intelligent connected vehicles, therefore, the set of collision-free paths There is Okay, representing There are several collision-free paths, each containing several path points at different times; For each path point in the set of collision-free paths, denoted as That is, the first The set of collision-free paths for vehicles at time steps The coordinates of the location in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. For time steps, The maximum number of steps; definition Indicating intelligent connected vehicles exist Collision-free paths within a time step, using Indicates all Intelligent connected vehicles in time step The set of path points; the set of collision-free paths Matches: Target location of intelligent connected vehicles The allocation matrix should correspond to the allocation scheme. The elements correspond to each other, and at each time step set of path points Neither should contain micro-level conflict relationships; First, utilize path planning The algorithm is based on the current allocation matrix. Perform path planning to obtain an initial set of collision-free paths. Check the current set of collision-free paths. Path points with micro-conflicts Add path points with micro-conflicts to the corresponding conflict point set. In the middle, the conflict points will be set As a constraint and utilizing path planning The algorithm executes the next round of path planning, repeatedly adding micro-conflicting waypoints and performing path planning again, until a set of collision-free paths is reached. When there are no more path points with microscopic conflicts in the path set, that is, when there are no more path points with microscopic conflicts in the path set, the path set is considered a collision-free path set. The process of eliminating microscopic conflicts has been completed, the iteration is now finished, and a collision-free path for each intelligent connected vehicle has been obtained. A3.3: Current Iteration Index Add 1; A3.4: The allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, is taken as the suboptimal allocation scheme for the current iteration. The lane-changing physical cost of the suboptimal allocation scheme is compared with the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme. If the lane-changing physical cost of the suboptimal allocation scheme is greater than the actual lane-changing physical cost of the collision-free path set corresponding to the optimal allocation scheme, then the collision-free path set corresponding to the optimal allocation scheme is taken as the optimal solution and A3.9 is executed; otherwise, A3.5 is executed. A3.5: Perform path planning on the suboptimal allocation scheme of the current iteration to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs; A3.6: Current Iteration Index Add 1; A3.7: Take the allocation scheme with the minimum lane-changing physical cost, excluding the allocation schemes that have already undergone path planning, as the suboptimal allocation scheme for the current iteration, perform path planning to obtain the corresponding set of collision-free paths and their actual lane-changing physical costs, and return to A3.6, until all allocation schemes are traversed and A3.8 is executed. A3.8: The set of collision-free paths that minimize the actual physical cost of lane changing is taken as the optimal solution; A3.9: Use the optimal solution as the final set of collision-free paths obtained from the solution.

6. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 1, characterized in that, The method of controlling intelligent connected vehicles to change lanes in the lane-changing zone and enter the car-following zone according to the collision-free path set, and planning the optimal acceleration trajectory at the micro level for each intelligent connected vehicle under the premise of satisfying car-following safety and physical kinematic constraints, specifically includes the following steps: C1: Intelligent connected vehicles enter the following zone; determine the entry time of each intelligent connected vehicle into the following zone. The initial speed when entering the car-following zone and the assigned target entry time into the intersection ; C2: Establish a decentralized energy-optimal control model for intelligent connected vehicles; For intelligent connected vehicles The decentralized energy optimal control model is established as follows: （13）； in, This is the input for vehicle control, namely acceleration; C3: Solve the decentralized energy-optimal control model to obtain the optimal acceleration trajectory at the micro level for each intelligent connected vehicle; To solve the decentralized energy-optimal control model, a Hamiltonian function is constructed using the Pontryagin maximum principle. for: （14）； in, For the position, speed, and acceleration of intelligent connected vehicles, and For costate variables, For the Lagrange multipliers corresponding to the constraints, For acceleration constraints, let represent the maximum and minimum acceleration values, respectively; For speed constraints, let represent the maximum and minimum speed values, respectively; According to the Euler-Lagrange equations, we can obtain: （15）； （16）； in, for The derivative, for The derivative, For the speed of intelligent connected vehicles; Let be a constant, and let it be... Integrating it yields ,in Let be the integration constant, and then according to the necessary condition for optimality: （17）； The optimal control law under conditions without velocity and acceleration constraints is derived to be a linear function of time. （18）； in, The optimal acceleration; Substituting this optimal acceleration into the vehicle dynamics equations and performing a quadratic integration yields the energy-optimal velocity at the microscopic level. With position Trajectory Programming Equation: （19）； （20）； in, It is the integration constant; The microscopic kinematic boundary conditions are set as follows: the initial velocity entering the catenary region. Initial position for entering the following zone Terminal location The boundary conditions for the costate variables are set as follows: , and Indicates the time when the assigned target enters the intersection. Costate variables under; A system of linear equations is constructed based on the boundary conditions of micro-kinematics and co-state variables, and the undetermined coefficients are solved in real time. : （21）； in, This is a transpose.

7. The method for scheduling connected vehicles at multi-lane unsignalized intersections considering lane changes according to claim 1, characterized in that, The process of planning the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, driving the intelligent connected vehicle to the stop line, and using the graph relationship method to achieve conflict-free passage in the scheduling area specifically includes the following steps: D1: Model all intelligent connected vehicles and construct a conflict directed graph, which includes several nodes, one node corresponds to one intelligent connected vehicle, and the edges represent the conflict relationships between intelligent connected vehicles. Define a conflict-directed graph The set of nodes in a conflicting directed graph is defined as follows: The set of edges of a conflicting directed graph It consists of the union of the set of one-way edges and the set of two-way edges: （22）； in, It is a set of one-way edges. It is a set of bidirectional edges; The one-way edges in the set of one-way edges represent car-following constraints, occurring between two connected vehicles entering the intersection from the same lane; if there exists a car-following constraint in the conflict directed graph from node... Pointing to node A one-way edge means that in physical space, the node Absolutely cannot compare to nodes First reach the conflict point at the intersection; （23）； in, The set of traffic splitting conflicts is the set of intelligent connected vehicles that have a car-following constraint with the intelligent connected vehicle. Represents an edge; The bidirectional edges in the bidirectional edge set represent cross-conflict constraints with a travel order game space. When there is a spatial intersection of the predetermined travel trajectories of intelligent connected vehicles in different directions or different lanes, if there is a bidirectional edge between two nodes, it means that the arrival order of the two intelligent connected vehicles can be interchanged, but they absolutely cannot occupy the conflict point at the same time. （24）； in, The cross-conflict set is the set of intelligent connected vehicles that have cross-conflicts with the intelligent connected vehicle. D2: The complement of the conflicting directed graph is taken as the coexisting undirected graph. By solving the maximum matching problem in the coexisting undirected graph, the vehicle pairing scheme with the highest spatiotemporal resource utilization at the intersection is obtained, so as to realize the conflict-free passage in the scheduling area. The coexisting undirected graph includes several vertices and edges. Each vertex represents an intelligent connected vehicle, and each edge represents that there is no conflict relationship between two vertices.

8. A networked vehicle dispatching system for multi-lane unsignalized intersections considering lane changes, used to implement the networked vehicle dispatching method for multi-lane unsignalized intersections considering lane changes as described in any one of claims 1-7, characterized in that, include: The area division module is used to determine the upstream road segment area of ​​an unsignalized intersection and construct the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection. It spatially divides the upstream road segment area of ​​the unsignalized intersection into two continuous sub-regions, namely the lane changing area and the following area. The mapping module is used to divide the road into discrete grid nodes in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, and to map the actual physical coordinates of the intelligent connected vehicle when it arrives at the starting point of the lane-changing area to the nearest discrete grid node in the relative coordinate system of the upstream road segment area of ​​the unsignalized intersection, so as to obtain the initial relative coordinates of each intelligent connected vehicle and determine the target lane of each intelligent connected vehicle. The lane change zone allocation scheme and collision-free path solution module is used to model the target lane allocation as a 0-1 integer allocation planning problem. By minimizing the physical cost of lane changing for the entire fleet, the 0-1 integer allocation planning problem is solved iteratively to obtain the allocation scheme and its corresponding physical cost for each iteration. The collision-free path set of the allocation scheme is solved in ascending order of physical cost of lane changing, and finally the collision-free path set with the minimum actual physical cost of lane changing is obtained. The optimal acceleration trajectory solution module for the car-following zone is used to control intelligent connected vehicles to change lanes in the lane-changing zone and enter the car-following zone according to the set of collision-free paths. Under the premise of satisfying car-following safety and physical kinematic constraints, it plans the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, including the acceleration of the intelligent connected vehicle at different times. The dispatch area passage module is used to plan the optimal acceleration trajectory at the micro level for each intelligent connected vehicle, so that the intelligent connected vehicle can drive to the stop line and achieve conflict-free passage in the dispatch area using the graph relationship method.

9. An electronic device, characterized in that, include: One or more processors, and a memory for storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the connected vehicle scheduling method for multi-lane unsignalized intersections considering lane changes as described in any one of claims 1-7.

10. A computer program product, characterized in that, Includes a computer program or instructions that, when executed by a processor, implement the multi-lane unsignalized intersection vehicle scheduling method considering lane changes as described in any one of claims 1-7.

Images