Cooperative scheduling method, device and equipment of autonomous driving vehicle, medium and product

Abstract

This application provides a method, apparatus, device, medium, and product for collaborative scheduling of autonomous vehicles, relating to the field of autonomous driving collaborative scheduling technology. The method includes: acquiring a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters; calculating the conflict area between any two vehicles based on the trajectory information and vehicle configuration parameters; constructing a vehicle conflict relationship graph based on the conflict areas and dividing it into at least one conflicting vehicle group; calculating the arrival time of each vehicle in each conflict area and the duration of passage through that conflict area; and using a preset partially greedy search algorithm, determining a target collaborative scheduling scheme based on the preset priority passage strategy, the arrival time of each vehicle in each conflict area, and the duration of passage through that conflict area, and generating corresponding scheduling instructions. This application, by combining vehicle future trajectory prediction, conflicting vehicle group decomposition, and multi-dimensional constraint scheduling optimization, can improve the scheduling accuracy, computational efficiency, and operational stability in multi-vehicle collaborative scenarios.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a method, apparatus, device, medium and product for the coordinated scheduling of autonomous vehicles. Background Technology

[0002] With the rapid development of autonomous driving technology, the demand for vehicle collaborative scheduling in closed or semi-closed scenarios such as ports, industrial parks, and mining areas is becoming increasingly urgent. In these scenarios, autonomous trucks and other vehicles need to complete high-density interactive operations in narrow roads, intersections, and work areas, with frequent behaviors such as path crossing, following other vehicles, and meeting oncoming traffic. This places extremely high demands on the real-time performance, accuracy, and stability of the scheduling system. Efficient collaborative scheduling technology has become a core support for ensuring operational safety and improving traffic efficiency.

[0003] Currently, vehicle cooperative scheduling technology mainly includes four types of solutions: First, traffic management based on traffic lights, which controls vehicle passage through fixed or adaptive traffic lights to adapt to fixed road scenarios; second, collision avoidance systems based on vehicle positioning, which trigger emergency braking when the distance between vehicles is too small based on real-time location information; third, scheduling methods based on predetermined routes, which pre-allocate fixed routes and passage time windows for vehicles; and fourth, priority scheduling methods based on current location, which determine the passage order based on the current status such as the distance between the vehicle and the conflict point.

[0004] However, making decisions based solely on the current location lacks the prediction and utilization of future vehicle trajectories, resulting in a lack of foresight in scheduling decisions. Relying on fixed grids or conflict points is prone to misjudging or omitting vehicle interaction relationships, and it fails to comprehensively consider constraints such as vehicle occupancy status, braking capacity, and intersection relationships, easily leading to unreasonable yielding relationships. In large-scale multi-vehicle scenarios, the computational complexity increases exponentially, failing to meet real-time requirements. Summary of the Invention

[0005] The collaborative scheduling method, apparatus, equipment, medium, and product for autonomous vehicles provided in this application are intended to address the problems of insufficient foresight, inaccurate conflict identification, and insufficient real-time performance in multi-vehicle scenarios in existing collaborative scheduling schemes.

[0006] In a first aspect, embodiments of this application provide a cooperative scheduling method for autonomous vehicles, the method comprising:

[0007] The system acquires a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters, wherein the trajectory information includes vehicle positioning parameters and planned trajectory parameters.

[0008] The conflict zone between any two vehicles is calculated based on the trajectory information and vehicle configuration parameters.

[0009] The vehicles are grouped according to the conflict area to obtain at least one conflict vehicle group, in which the vehicles in the conflict vehicle group have direct or indirect conflict relationships with each other.

[0010] Calculate the time it takes for each vehicle to arrive at each conflict zone and the duration it takes to pass through that conflict zone;

[0011] For each conflicting vehicle group, a pre-set partial greedy search algorithm is used to determine the target coordinated scheduling scheme based on the pre-set priority passage strategy, the arrival time of each vehicle in each conflict area, and the duration of passage through the conflict area, so that each vehicle travels according to the target coordinated scheduling scheme.

[0012] In one possible implementation, the vehicle configuration parameters include vehicle length, width, and track width; the calculation of the conflict area between any two vehicles based on the trajectory information and the vehicle configuration parameters includes:

[0013] The trajectory lines and curvature curves corresponding to each vehicle are generated by fitting each trajectory point in the planned trajectory.

[0014] Calculate the vehicle bounding box corresponding to each trajectory point based on the vehicle configuration parameters, trajectory lines, and curvature curves;

[0015] The vehicle bounding boxes corresponding to each trajectory point are combined to generate the trajectory geometric region corresponding to the vehicle. The trajectory geometric region is the spatial occupancy range of the vehicle within a preset time range.

[0016] Calculate the spatial overlap of the geometric regions of the trajectories of any two vehicles;

[0017] The vehicle following relationship is determined based on the spatial overlap using a preset following determination rule;

[0018] The overlapping spatial portions that satisfy the vehicle following relationship are deleted, and the boundary points of the remaining overlapping spatial portions are projected onto the corresponding trajectory lines to obtain the conflict area.

[0019] In one possible implementation, calculating the arrival time of each vehicle at each conflict zone and the duration of passage through the conflict zone includes:

[0020] Calculate the expected speed at each trajectory point based on the vehicle's current positioning parameters, road speed limit, vehicle acceleration limit, and curvature curve;

[0021] The average speed algorithm is used to calculate the time to reach each trajectory point based on the expected speed of each trajectory point, so as to obtain the time for the vehicle to reach each conflict area.

[0022] The time required to pass through each conflict zone is determined based on the arrival time at each trajectory point.

[0023] In one possible implementation, the step of using a preset partially greedy search algorithm to determine a target cooperative scheduling scheme based on the preset priority passage strategy, the arrival time of each vehicle in each conflict zone, and the duration of passage through the conflict zone includes:

[0024] The conflict areas of each vehicle in the conflict group are merged, and conflict areas that completely contain, cover, or highly overlap are deleted to obtain updated conflict areas.

[0025] Construct a search tree and determine the vehicle passage order in each updated conflict area according to the order of updating the conflict areas to obtain multiple candidate cooperative scheduling schemes;

[0026] The total cost corresponding to the complete passage sequence of each candidate cooperative scheduling scheme is calculated using a preset cost function;

[0027] The candidate coordinated scheduling scheme with the lowest total cost is determined as the target coordinated scheduling scheme for the current conflicting train crews.

[0028] In one possible implementation, after determining the target cooperative scheduling scheme, the method further includes:

[0029] Extract the vehicles that need to give way and the conflict areas that need to give way from the target collaborative scheduling scheme;

[0030] Calculate the trajectory coordinates of the stopping point based on the entrance location of the conflict zone that needs to be yielded to and the preset safety distance;

[0031] The trajectory coordinates of the stopping point are projected onto the trajectory line to generate the stopping point;

[0032] If the conflict area satisfies the vehicle following relationship, no parking point will be generated.

[0033] In one possible implementation, it also includes:

[0034] Generate scheduling instructions based on the target collaborative scheduling scheme;

[0035] The dispatch instructions and stop point information are encapsulated into messages and sent to each vehicle through the communication link of the cloud dispatch service, so that each vehicle can drive according to the target collaborative dispatch plan.

[0036] In one possible implementation, it also includes:

[0037] Determine whether the conflict area calculated in the current cycle is consistent with that in the previous cycle;

[0038] Determine whether the vehicles and conflict relationships included in each conflicting vehicle group in the current cycle are consistent with those in the previous cycle;

[0039] Determine whether the target coordinated scheduling scheme of the previous period is in the set of candidate coordinated scheduling schemes for the current period;

[0040] Determine whether the difference between the total cost of the target coordinated scheduling scheme in the previous period and the total cost of the target coordinated scheduling scheme in the current period is less than a preset cost threshold;

[0041] If all the above conditions are met, the scheduling instruction of the previous cycle will be determined as the scheduling instruction of the current cycle.

[0042] Secondly, embodiments of this application provide a cooperative scheduling device for autonomous vehicles, comprising:

[0043] The acquisition module is used to acquire a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters. The trajectory information includes vehicle positioning parameters and planned trajectory parameters.

[0044] The calculation module is used to calculate the conflict area between any two vehicles based on the trajectory information and vehicle configuration parameters.

[0045] The grouping module is used to group the vehicles according to the conflict area to obtain at least one conflict vehicle group, wherein the vehicles in the conflict vehicle group have a direct or indirect conflict relationship with each other.

[0046] The calculation module is also used to calculate the time it takes for each vehicle to arrive at each conflict zone and the duration of passing through the conflict zone.

[0047] The determination module is used to determine the target coordinated scheduling scheme for each conflicting vehicle group by using a preset partial greedy search algorithm based on the preset priority passage strategy, the time of each vehicle to arrive at each conflict area and the duration of passing through the conflict area, so that each vehicle travels according to the target coordinated scheduling scheme.

[0048] Thirdly, embodiments of this application provide a collaborative scheduling device for autonomous vehicles, including: a memory and a processor;

[0049] The memory stores computer-executed instructions;

[0050] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0051] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0052] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0053] The collaborative scheduling method, device, equipment, medium, and product for autonomous vehicles provided in this application's embodiments pre-set a priority passage strategy. By comprehensively considering actual constraints such as vehicle occupancy status, braking capacity, and intersection interaction relationships, a priority passage relationship conforming to traffic rules and scenarios is constructed. This clarifies the passage priority of different vehicles in case of conflict, facilitating the rational allocation of road resources and avoiding disorderly competition. By comprehensively considering vehicle trajectory information and configuration parameters, the potential conflict area between any two vehicles can be accurately calculated. Precise conflict area calculation makes subsequent group processing and collaborative scheduling more targeted, avoiding unnecessary global scheduling, reducing computational load and scheduling complexity, and improving scheduling efficiency and accuracy. Grouping vehicles with conflict relationships into conflict vehicle groups transforms the complex multi-vehicle conflict problem into multiple relatively independent problems within vehicle groups, reducing computational dimensionality and complexity, and effectively improving real-time performance. Calculating the arrival time and passage duration of each vehicle to the conflict area allows for estimation of the actual arrival time of vehicles, preventing unreasonable scheduling situations that violate traffic efficiency, such as distant vehicles yielding to nearby vehicles or slower vehicles yielding to faster vehicles. Duration information also helps optimize scheduling schemes, improve road traffic efficiency, and reduce vehicle waiting time. The pre-defined partial greedy search algorithm combines the efficiency of greedy algorithms with the comprehensiveness of partial search, enabling it to find a relatively optimized cooperative scheduling scheme in a short time. This avoids the enormous computational burden of global search and improves the efficiency of generating scheduling schemes. The target cooperative scheduling scheme determined by this algorithm ensures that vehicles pass through conflict areas in a reasonable order and at the right time, avoiding collisions and ensuring driving safety. It also makes full use of road resources, reduces vehicle waiting and delays, and enables cooperative and orderly driving of autonomous vehicles. Attached Figure Description

[0054] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0055] Figure 1 An application scenario diagram of the cooperative scheduling method for autonomous vehicles provided in this application;

[0056] Figure 2 A flowchart illustrating a collaborative scheduling method for autonomous vehicles provided in an embodiment of this application;

[0057] Figure 3 A flowchart illustrating a collaborative scheduling method for autonomous vehicles provided in another embodiment of this application;

[0058] Figure 4A schematic diagram of the structure of a collaborative scheduling device for an autonomous vehicle provided in an embodiment of this application;

[0059] Figure 5 This is a schematic diagram of the structure of a collaborative scheduling device for autonomous vehicles provided in an embodiment of this application.

[0060] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0061] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0062] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0063] To clearly understand the technical solution of this application, the solutions of the prior art will be described in detail first.

[0064] With the continuous evolution of autonomous driving technology, the demand for collaborative scheduling of autonomous vehicles is becoming increasingly urgent in closed and semi-closed scenarios such as ports, industrial parks, and mining areas. In these scenarios, autonomous trucks and other vehicles need to conduct high-density interactive operations on narrow roads, intersections, and work areas, with frequent occurrences of path crossings and following other vehicles. This places stringent requirements on the real-time performance, accuracy, and stability of the scheduling system, making efficient collaborative scheduling technology a key support for ensuring operational safety and improving traffic efficiency. Currently, vehicle collaborative scheduling technology mainly includes four mainstream solutions: First, traffic management based on traffic lights, which uses fixed or adaptive traffic lights to control vehicle passage and adapts to fixed road scenarios; second, collision avoidance systems based on vehicle positioning, which rely on real-time location information to trigger emergency braking when the distance between vehicles is too small; third, scheduling methods based on predetermined paths, which pre-plan fixed paths for vehicles and allocate passage time windows; and fourth, priority scheduling methods based on current location, which determine the passage order based on real-time status such as the distance between the vehicle and the conflict point. However, existing technologies have many shortcomings. Scheduling decisions are based solely on the current location of vehicles without considering future trajectory predictions, resulting in a lack of foresight in decision-making. Conflict area identification relies on fixed grids or conflict points, which can easily lead to misjudgments or omissions in interaction relationships. Constraints such as vehicle occupancy status, braking capacity, and intersection relationships are not comprehensively considered, which can easily lead to unreasonable yielding schemes. In large-scale multi-vehicle scenarios, computational complexity increases exponentially, making it difficult to meet real-time scheduling requirements.

[0065] Therefore, to address the problems of lack of foresight in scheduling decisions, inaccurate conflict identification, and insufficient real-time performance in large-scale multi-vehicle scenarios in existing technologies, this application proposes a vehicle cooperative scheduling scheme based on trajectory prediction. The scheme first obtains vehicle positioning parameters, planned trajectory parameters, and vehicle configuration parameters to geometrically model the future space occupancy of vehicles. It then calculates the overlap relationship between the geometric regions of any two vehicle trajectories, filters following scenarios, and projects the trajectory coordinate range to achieve accurate identification of conflict areas. Secondly, it constructs a conflict relationship graph based on direct or indirect conflict relationships between vehicles and divides them into conflict vehicle groups, decomposing the global scheduling problem into multiple relatively independent sub-problems. Finally, based on the time vehicles arrive at the conflict area, the duration of passage through the conflict area, and hierarchical priority passage rules, it uses a combination of partial greedy search and pruning to determine the target cooperative scheduling scheme, thereby improving the real-time performance and stability of multi-vehicle cooperative scheduling while ensuring traffic safety.

[0066] Figure 1 This is an application scenario diagram illustrating the cooperative scheduling method for autonomous vehicles provided in this application, such as... Figure 1As shown, the scenario diagram corresponding to the collaborative scheduling method for autonomous vehicles provided in this application includes: vehicle A 101, vehicle B 102, vehicle C 103, vehicle D 104, vehicle E 105, and server 106. The collaborative scheduling device for autonomous vehicles can be integrated into server 106.

[0067] Understandably, server 106 is configured with a database for storing preset priority access policies, which can be dynamically adjusted according to actual conditions.

[0068] Specifically, server 106 acquires the trajectory information and vehicle configuration parameters of vehicles A 101, B 102, C 103, D 104, and E 105 in real time via existing communication links. Then, server 106 calculates the conflict zone between any two vehicles based on the trajectory information and vehicle configuration parameters, and groups the vehicles according to the conflict zones to obtain at least one conflicting vehicle group. Server 106 calculates the arrival time and transit time of each vehicle in each conflict zone, and uses a pre-defined partially greedy search algorithm for each conflicting vehicle group. Based on a pre-defined priority passage strategy, the arrival time and transit time of each vehicle in each conflict zone, server 106 determines the target collaborative scheduling scheme. Finally, server 106 converts the target collaborative scheduling scheme into scheduling instructions and stop point information, and sends them to the planning and control modules of the corresponding vehicles via existing communication links, so that each vehicle travels according to the target collaborative scheduling scheme.

[0069] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0070] Figure 2 This is a flowchart illustrating a cooperative scheduling method for autonomous vehicles provided in an embodiment of this application, as shown below. Figure 2 As shown, the execution entity in this embodiment is a collaborative scheduling device for autonomous vehicles. This device can be implemented through a computer program, or through a medium storing the relevant computer program, such as a USB flash drive and / or optical disc, or through a physical device integrating or installing the relevant computer program, such as a chip or a collaborative scheduling device for autonomous vehicles. The collaborative scheduling device for autonomous vehicles can be a server, a server cluster, etc. The collaborative scheduling method for autonomous vehicles provided in this embodiment includes the following steps:

[0071] S201. Obtain the preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters. The trajectory information includes the vehicle's positioning parameters and planned trajectory parameters.

[0072] Among them, the preset priority passage strategy refers to a set of multi-layered rules that are pre-set to determine the priority of vehicles in conflict areas.

[0073] Trajectory information refers to the real-time motion status and future driving path related data of autonomous vehicles, including real-time dynamic data such as the vehicle's current position coordinates, orientation angle, and driving speed; and future trajectory point sequences generated by the vehicle based on the navigation path and driving intention.

[0074] Among them, vehicle configuration parameters refer to the inherent geometric and performance parameters of a vehicle, mainly including body length, width, track width, overhang dimensions, etc.

[0075] Understandably, it's necessary to pre-analyze vehicle traffic rules for closed or semi-closed scenarios and construct a hierarchical priority traffic strategy. This strategy comprehensively considers factors from high to low priority, such as whether a vehicle has already occupied a conflict zone or is in a following state, whether the vehicle cannot stop safely due to insufficient braking distance, whether the vehicle has abnormal parking, the proximity of the vehicle to the conflict zone, the scheduling requirements of the previous cycle, and intersection traffic rules. These rules are then transformed into computable decision conditions and stored in the system strategy library for use during the scheduling process.

[0076] For example, vehicles that have entered the conflict zone or formed a stable following relationship can pass first; when the braking distance corresponding to the vehicle's current speed and maximum deceleration is greater than the remaining distance to the conflict zone, it can be determined that it is difficult to stop safely in time and pass first; the conflict relationship of abnormally stopped vehicles can trigger reverse priority allocation; when a vehicle is too close to the conflict zone, it can be given higher priority to avoid frequent switching; if a scheduling solution that can be continued has been formed in the previous cycle, the existing passage order can be maintained first; in the intersection scenario, priority passage can also be determined by combining whether the vehicle has entered the intersection, the road attributes, and the turning relationship.

[0077] Specifically, the cloud-based V2V Server collects the positioning parameters of each autonomous vehicle participating in the scheduling in real time through existing communication links. Simultaneously, it retrieves the path planning modules of each vehicle to obtain its planned trajectory parameters, forming a complete set of trajectory information including real-time status and future paths. Configuration parameters for each vehicle, including key geometric data such as body length, width, wheelbase, and overhang dimensions, are extracted from the vehicle's inherent attribute database. These parameters can be pre-entered into the database, and the scheduling system can directly read and access them.

[0078] S202. Calculate the conflict area between any two vehicles based on trajectory information and vehicle configuration parameters.

[0079] The conflict zone refers to the area where the geometric regions of the future trajectories of two autonomous vehicles overlap, and it is a key area where traffic conflicts may occur during the driving of the two vehicles.

[0080] Specifically, based on the acquired vehicle configuration parameters, including vehicle length, width, track width, and overhang dimensions, a corresponding vehicle bounding box is calculated for each trajectory point on the planned trajectory. The vehicle bounding boxes of all trajectory points are combined in trajectory order to form the geometric region of the vehicle's future trajectory. For any two autonomous vehicles, spatial overlap detection is performed on their trajectory geometric regions. If overlap exists, a pre-defined following rule is used to determine whether it belongs to a same-direction following scenario. Overlapping portions that satisfy the following relationship are not treated as conflict areas; for the remaining overlapping portions, their boundary points are projected onto the trajectory lines of the corresponding vehicles to obtain the conflict range in their respective trajectory coordinate systems, and a conflict region between the two vehicles is generated accordingly.

[0081] S203. The vehicles are grouped according to the conflict area to obtain at least one conflict vehicle group, in which the vehicles in the conflict vehicle group have direct or indirect conflict relationships with each other.

[0082] A conflict group refers to a set of vehicles that have a direct or indirect conflict relationship. Any vehicle in the group has a conflict area with at least one other vehicle, or has an indirect conflict relationship with other vehicles in the group.

[0083] Specifically, each autonomous vehicle participating in the scheduling is treated as a vertex in the conflict graph. Then, all vehicles are paired up; if a valid conflict area has been calculated between two vehicles (i.e., a direct conflict relationship exists), an undirected edge is created between the vertices corresponding to these two vehicles. This ultimately forms an undirected conflict graph with vehicles as vertices and direct conflict relationships as edges. An adjacency list data structure is used for storage, such as graph[vehicle A] = {conflicting vehicles B, C, ...}.

[0084] Furthermore, a breadth-first search algorithm is used to traverse and decompose the conflict graph. Starting from any unvisited vehicle vertex, all vertices directly connected to that vertex are traversed, and then the associated vertices of these connected vertices are traversed until no new associated vertices are found. These interconnected vertex sets constitute a connected component, i.e., a conflicting vehicle group. This traversal process is repeated until all vehicle vertices have been visited, ultimately dividing all vehicles into at least one conflicting vehicle group.

[0085] Furthermore, each conflicting vehicle group is tagged with attributes. If a conflicting vehicle group contains only one vehicle, meaning that vehicle has no direct or indirect conflict with any other vehicle, it is classified as a single-vehicle conflicting vehicle group and does not require further scheduling optimization. If a conflicting vehicle group contains two or more vehicles, it is classified as a multi-vehicle conflicting vehicle group and must be considered for scheduling optimization.

[0086] S204. Calculate the time it takes for each vehicle to arrive at each conflict zone and the duration of passage through that conflict zone.

[0087] The arrival time in the conflict zone refers to the expected time when an autonomous vehicle, starting from its current location and following its planned trajectory, will arrive at the entrance to the conflict zone.

[0088] The time taken to pass through the conflict zone refers to the expected time required for a vehicle to travel from the entrance of the conflict zone to the exit of the conflict zone.

[0089] Understandably, the initial state parameters of the vehicle are first determined, including the current speed and the corresponding point of the current position on the planned trajectory. At the same time, the full trajectory point data of the vehicle's planned trajectory, the speed limit information of each road segment, the vehicle's own acceleration limit, and the trajectory curvature curve are retrieved.

[0090] Specifically, based on the vehicle's current position in the trajectory projection coordinate system, the system iterates forward along the planned trajectory from the trajectory point corresponding to the current position. For each trajectory point, it determines the road segment speed limit constraint, the speed constraint based on the vehicle's acceleration and deceleration capabilities, and the safe speed constraint based on the trajectory curvature, and takes the smaller value as the expected speed of the trajectory point. Then, based on the trajectory distance between adjacent trajectory points and the speed of the preceding and following trajectory points, it estimates the time increment, and accumulates the estimated arrival time of each trajectory point point by point.

[0091] Furthermore, based on the entrance and exit positions of the conflict zone on the trajectory line, the corresponding positions are projected onto the vehicle trajectory coordinates, and the estimated arrival times of the entrance and exit positions are obtained through interpolation or querying. The estimated arrival time of the entrance position is used as the time for the vehicle to arrive at the conflict zone, and the difference between the corresponding times of the exit and entrance is used as the duration for the vehicle to pass through the conflict zone.

[0092] S205. For each conflicting vehicle group, a pre-set partial greedy search algorithm is used to determine the target coordinated scheduling scheme based on the pre-set priority passage strategy, the time for each vehicle to arrive at each conflict area, and the duration of passing through the conflict area, so that each vehicle travels according to the target coordinated scheduling scheme.

[0093] Among them, the pre-defined greedy search algorithm is a heuristic optimization algorithm that combines a greedy strategy with search tree traversal. Its core is to determine the vehicle passage order for each conflict area under the premise of satisfying priority constraints, while retaining low-cost candidate solutions through pruning operations, thus achieving a balance between scheduling optimality and computational real-time performance.

[0094] Among them, the target collaborative scheduling scheme refers to the optimal vehicle passage sequence scheme generated for each conflicting vehicle group. This scheme satisfies all priority passage constraints and has the minimum corresponding total cost, which can directly guide vehicles to pass or give way.

[0095] Understandably, for each conflicting vehicle group, the first step is to extract all conflict areas traversed by each vehicle within the group, as well as the trajectory range of each vehicle within the corresponding conflict area. Next, conflict areas that completely encompass, cover, or highly overlap are merged to reduce the number of optimization variables. Finally, a complete set of optimization parameters is compiled, including a list of conflict areas for each vehicle, the earliest arrival time in each conflict area, the time taken to pass through each conflict area, and a priority-directed graph generated based on the priority passage strategy.

[0096] Specifically, a search tree is constructed using the pending scheduling status of conflicting vehicle groups as the root node. Each level of the tree corresponds to a conflict region or updates the vehicle passage order within a conflict region. The search tree is expanded layer by layer using a level-order traversal. During each expansion, the local scheduling solution information of the parent node is preserved, and vehicles within the current conflict region that satisfy priority constraints and have an in-degree of 0 are selected for sorting. For each vehicle that meets the conditions, its corresponding child nodes are expanded, the vehicle is added to the current passage sequence, and the in-degree of other vehicles awaiting sorting in that conflict region is updated. This process continues until the sorting of the current conflict region is completed, before moving on to the next conflict region.

[0097] Optionally, during the search tree expansion process, if the number of nodes at a certain level exceeds a preset threshold, a pruning operation is initiated. For each candidate node at that level, the corresponding cost is calculated based on the currently formed partial passage sequence. Several candidate nodes with lower costs and meeting the feasibility requirements are retained, while nodes with higher costs or that are not feasible are removed. The cost can comprehensively consider the passage completion time distribution of each vehicle and the vehicle weight, in order to balance overall passage efficiency with the scheduling needs of key vehicles.

[0098] Furthermore, after the search tree has traversed all conflict zones and sorted the passage, the leaf nodes of the tree represent the complete candidate scheduling schemes for the conflicting vehicle group. The total cost corresponding to each leaf node is calculated, and the leaf node with the minimum total cost is selected. The vehicle passage order corresponding to this node is the target cooperative scheduling scheme for the conflicting vehicle group.

[0099] The collaborative scheduling method for autonomous vehicles provided in this application pre-sets a priority passage strategy. It comprehensively considers actual constraints such as vehicle occupancy status, braking capacity, and intersection interaction relationships to construct a priority passage relationship that conforms to traffic rules and scenarios. This clarifies the passage priority of different vehicles in case of conflict, facilitating the rational allocation of road resources and avoiding disorderly competition. By comprehensively considering vehicle trajectory information and configuration parameters, the area where any two vehicles may clash can be accurately calculated. Precise conflict area calculation makes subsequent group processing and collaborative scheduling more targeted, avoiding unnecessary global scheduling, reducing computational load and scheduling complexity, and improving scheduling efficiency and accuracy. Grouping vehicles with conflict relationships into conflict vehicle groups transforms the complex multi-vehicle conflict problem into multiple relatively independent problems within vehicle groups, reducing computational dimensionality and complexity, and effectively improving real-time performance. Calculating the arrival time and passage duration of each vehicle to the conflict area allows for estimation of the actual arrival time of vehicles, preventing unreasonable scheduling situations that violate traffic efficiency, such as distant vehicles yielding to nearby vehicles or slower vehicles yielding to faster vehicles. The duration information also helps optimize the scheduling scheme, improve road traffic efficiency, and reduce vehicle waiting time. The pre-defined partial greedy search algorithm combines the efficiency of greedy algorithms with the comprehensiveness of partial search, enabling it to find a relatively optimized cooperative scheduling scheme in a short time. This avoids the enormous computational burden of global search and improves the efficiency of generating scheduling schemes. The target cooperative scheduling scheme determined by this algorithm ensures that vehicles pass through conflict areas in a reasonable order and at the right time, avoiding collisions and ensuring driving safety. It also makes full use of road resources, reduces vehicle waiting and delays, and enables cooperative and orderly driving of autonomous vehicles.

[0100] As an optional implementation, based on the above embodiments, the vehicle configuration parameters include vehicle length, width, and track width; the conflict area between any two vehicles is calculated based on trajectory information and vehicle configuration parameters, including:

[0101] The trajectory lines and curvature curves corresponding to each vehicle are generated by fitting each trajectory point in the planned trajectory.

[0102] Calculate the vehicle bounding box corresponding to each trajectory point based on vehicle configuration parameters, trajectory lines, and curvature curves;

[0103] The vehicle bounding boxes corresponding to each trajectory point are combined to generate the trajectory geometric region corresponding to the vehicle. The trajectory geometric region is the spatial occupancy range of the vehicle within a preset time range.

[0104] Calculate the spatial overlap of the geometric regions of the trajectories of any two vehicles;

[0105] The vehicle following relationship is determined based on the spatial overlap using a preset following determination rule;

[0106] The overlapping spatial portions that satisfy the vehicle following relationship are deleted, and the boundary points of the remaining overlapping spatial portions are projected onto the corresponding trajectory lines to obtain the conflict area.

[0107] The vehicle bounding box refers to the rectangular frame constructed for each trajectory point on the planned trajectory based on the vehicle's configuration parameters. This rectangular frame can completely enclose the vehicle body at the corresponding location, accurately representing the vehicle's spatial occupancy and position at that trajectory point.

[0108] The trajectory geometry region refers to the spatial area formed by sequentially combining the vehicle bounding boxes of all trajectory points on the planned trajectory. This region reflects the total space actually occupied by the vehicle while traveling along the planned trajectory within a preset time range.

[0109] The curvature curve refers to a curve generated by calculating the degree of curvature at each point on the trajectory, based on the trajectory line corresponding to the geometric region of the trajectory. This curve can be used to query the curvature at any position on the trajectory.

[0110] Among them, the following relationship refers to the driving state in which two vehicles travel in the same direction along basically overlapping trajectories and meet the preset judgment conditions.

[0111] Specifically, for the planned trajectory of a single autonomous vehicle, the position and orientation information of all discrete trajectory points on the trajectory are extracted. Based on the spatial distribution of the trajectory points, a continuous and smooth vehicle trajectory line is fitted using linear interpolation. This trajectory line serves as the centerline for the vehicle's planned driving. Simultaneously, the curvature value of each discrete point on the trajectory line is calculated, and the discrete curvature values ​​are then fitted into a continuous curvature curve through interpolation.

[0112] Furthermore, for each trajectory point on the planned trajectory, the spatial orientation of the vehicle at that location is first determined based on its position, orientation, and curvature value corresponding to the curvature curve. Then, using this trajectory point as a reference point, a rectangular frame that completely encloses the vehicle body in that orientation is constructed according to parameters such as vehicle length and width; this is the vehicle bounding box corresponding to that trajectory point. The vehicle bounding boxes corresponding to all trajectory points on the planned trajectory are then spatially combined according to the chronological order of the trajectory to form a continuous spatial region—the geometric region of the trajectory that the vehicle will sweep through in the future.

[0113] Furthermore, for any two autonomous vehicles requiring detection, their trajectory geometric regions are extracted. A geometric overlap detection method is used to spatially compare the two trajectory geometric regions to determine if they intersect. If an intersection exists, the spatial range of that intersection is extracted, i.e., the spatial overlap portion of the two vehicles' trajectory geometric regions.

[0114] Optionally, the preset following determination rules can be: the preceding vehicle has entered the conflict zone; the preceding vehicle is ahead of the following vehicle by at least one vehicle length along the trajectory direction; the vertical distance from the rear axle point of the preceding vehicle to the planned trajectory line of the following vehicle is less than a preset threshold, where the preset threshold is usually set to half the width of the vehicle; and the angle between the current orientation of the preceding vehicle and the tangent direction of the trajectory of the following vehicle is less than a preset angle threshold.

[0115] Furthermore, if multiple conditions in the preset following relationship determination rules are met simultaneously, a following relationship between two vehicles is determined. If the two vehicles are determined to be following each other, the corresponding spatial overlap is deleted, and this part is not considered a conflict area. If they are not determined to be following each other, the corresponding spatial overlap is retained. The boundary points of the overlap area are projected onto their respective trajectory lines to obtain the conflict range s_range = (s_min, s_max) in their respective trajectory coordinate systems, which is the final conflict area. The conflict area information is recorded, including the vehicle identifiers of the two vehicles involved in the conflict and the conflict range.

[0116] The collaborative scheduling method for autonomous vehicles provided in this application calculates the vehicle bounding boxes corresponding to each trajectory point in the planned trajectory based on vehicle configuration parameters such as vehicle length, width, and wheelbase. This accurately depicts the spatial range of the vehicle at each trajectory point, avoiding errors caused by simplification to a point mass or simple graphic, and closely reflects reality, laying the foundation for accurate identification of conflict areas. By combining the vehicle bounding boxes corresponding to each trajectory point, a vehicle trajectory geometric region is generated, covering its entire spatial occupancy within a preset time period, allowing for a more comprehensive capture of areas that may conflict with other vehicles. The spatial overlap of any two vehicle trajectory geometric regions is calculated, and the following relationship is determined using preset following rules. Overlapping portions that satisfy the following relationship are deleted, eliminating spatial overlap caused by normal following, and retaining portions with genuine conflict risk, significantly improving the accuracy of conflict identification. Precise conflict areas provide crucial information for collaborative scheduling, helping to further improve the accuracy of collaborative scheduling.

[0117] As an optional implementation, based on the above embodiments, the time for each vehicle to arrive at each conflict zone and the duration of passage through the conflict zone are calculated, including:

[0118] Calculate the expected speed at each trajectory point based on the vehicle's current positioning parameters, road speed limit, vehicle acceleration limit, and curvature curve;

[0119] An average speed algorithm is used to calculate the time to reach each trajectory point based on the expected speed of each trajectory point, so as to obtain the time for the vehicle to reach each conflict area.

[0120] The time required to pass through each conflict zone is determined based on the arrival time at each trajectory point.

[0121] Among them, the speed limit for a road segment refers to the maximum permissible speed for autonomous vehicles on a road segment.

[0122] Among them, the vehicle acceleration limit refers to the maximum acceleration and deceleration threshold determined by the vehicle's own dynamic performance and safety control strategy.

[0123] The expected speed refers to the safe and feasible driving speed of a vehicle at a certain point on the planned trajectory, after taking into account the speed limit of the road section, acceleration constraints, and trajectory curvature constraints.

[0124] Specifically, the current positioning parameters of the vehicle are obtained to determine the trajectory point corresponding to the vehicle's current position, which serves as the starting point for speed calculation. Then, subsequent trajectory points are traversed one by one along the planned trajectory. For each trajectory point, the speed limit value of the road segment where that trajectory point is located is retrieved. Based on the vehicle's current speed and acceleration limit, the acceleration constraint speed for that trajectory point is calculated. Then, based on the curvature curve value corresponding to that trajectory point, the curvature constraint speed to ensure vehicle turning safety is determined. The minimum value among these three speeds is taken as the expected speed for that trajectory point, and this expected speed is updated as the calculation reference speed for the next trajectory point.

[0125] Furthermore, the vehicle's current time is initialized to 0, and its current speed is the initial speed. For two adjacent trajectory points, the displacement distance Δs = s[i] - s[i-1] between the two points is first calculated; then, based on the speed of the previous trajectory point and the expected speed of the current trajectory point, the average speed between the two points v_next = min(speed limit, acceleration constraint speed, curvature constraint speed) is calculated. Next, the displacement distance is divided by the average speed to obtain the time increment for the vehicle to pass through the road segment. This time increment is added to the arrival time of the previous trajectory point to obtain the arrival time of the current trajectory point. This method is used to calculate the arrival times of all trajectory points.

[0126] Furthermore, based on the entrance and exit locations of the conflict zones on the trajectory line, the arrival time of the vehicle at the entrance of the conflict zone is determined by interpolation, which is the time for the vehicle to arrive at each conflict zone. Based on the range of the conflict zone on the trajectory line, the arrival times of the trajectory points corresponding to the entrance and exit of the conflict zone are extracted. The difference between the exit arrival time and the entrance arrival time is the duration of the vehicle's passage through the conflict zone.

[0127] The collaborative scheduling method for autonomous vehicles provided in this application calculates the expected speed of each trajectory point based on the vehicle's current positioning parameters, road speed limits, acceleration limits, and curvature curves. Then, an average speed algorithm is used to determine the arrival time at each trajectory point, thereby determining the arrival time in the conflict zone. This multi-factor calculation method accurately predicts the vehicle's arrival time in the conflict zone, providing a basis for early detection of potential conflicts and allowing the system sufficient time to formulate response strategies to avoid conflicts. Simultaneously, accurately calculating the time to pass through the conflict zone facilitates the rational planning of vehicle travel time in that area. The scheduling system can then arrange for vehicles to pass through in an orderly manner, preventing vehicles from lingering or causing congestion in the conflict zone and improving traffic flow. The calculation process comprehensively considers factors such as vehicle status, road conditions, and the vehicle's own performance, resulting in more realistic results. The scheduling scheme formulated based on this is better adapted to different vehicles and road environments, improving the scientific and rational nature of the scheduling.

[0128] As an optional implementation, based on the above embodiments, a pre-defined greedy search algorithm is used to determine the target cooperative scheduling scheme based on a pre-defined priority passage strategy, the arrival time of each vehicle in each conflict zone, and the duration of passage through the conflict zone, including:

[0129] The conflict areas of each vehicle in the conflict group are merged, and conflict areas that completely contain, cover, or highly overlap are deleted to obtain updated conflict areas.

[0130] Construct a search tree and determine the vehicle passage order in each updated conflict area according to the order of updating the conflict areas to obtain multiple candidate cooperative scheduling schemes;

[0131] The total cost corresponding to the complete passage sequence of each candidate cooperative scheduling scheme is calculated using a preset cost function;

[0132] The candidate coordinated scheduling scheme with the lowest total cost is determined as the target coordinated scheduling scheme for the current conflicting train crews.

[0133] Among them, the updated conflict area refers to the set of areas obtained after merging the original conflict areas within the conflicting train group.

[0134] Among them, the candidate collaborative scheduling scheme refers to the vehicle passage order scheme generated during the search tree traversal process that satisfies the preset priority passage strategy. Each scheme corresponds to a leaf node of the search tree.

[0135] The preset cost function refers to the calculation rules used to measure the quality of the scheduling scheme. The total cost is the sum of the costs of all individual vehicles in the conflicting vehicle group. The cost of an individual vehicle is equal to the total time it takes for the vehicle to pass through all conflict areas multiplied by the vehicle weight. The vehicle weight will be dynamically adjusted according to the historical yielding status.

[0136] Specifically, for a single conflicting vehicle group, all original conflict areas that each vehicle in the group needs to pass through are extracted. Spatial correlation analysis is performed on these conflict areas to identify areas that completely contain, cover, or highly overlap, and these areas are merged into a unified updated conflict area. Using the initial scheduling state of the conflicting vehicle group as the root node, a search tree is constructed, with each level of the tree corresponding to a vehicle passage sorting task for an updated conflict area. The search tree is expanded layer by layer according to the order of updating conflict areas. During expansion, the local scheduling solution information of the current node is first preserved, and then, based on a preset priority passage strategy, vehicles with an in-degree of 0 in the current conflict area are selected. One vehicle is randomly selected from those that meet the criteria and added to the current passage sequence, generating a child node of the current node. Then, the in-degree of other vehicles to be sorted in this conflict area is updated, and this operation is repeated until all vehicles in the current conflict area have completed their passage sorting, before proceeding to the next updated conflict area sorting process. After all updated conflict areas have been sorted, the leaf nodes of the search tree represent multiple candidate cooperative scheduling schemes.

[0137] Furthermore, based on a preset cost function, the total cost is calculated for each candidate coordinated scheduling scheme. During the calculation, the total time for each vehicle to traverse all conflict zones within the scheme is first calculated, then multiplied by the vehicle's dynamic weight to obtain the cost per vehicle. Finally, the costs of all individual vehicles are summed to obtain the total cost of the scheme. After calculating the costs of all candidate schemes, the total cost values ​​are compared, and the candidate coordinated scheduling scheme with the lowest total cost is determined as the target coordinated scheduling scheme for the current conflicting vehicle group.

[0138] The collaborative scheduling method for autonomous vehicles provided in this application merges conflict areas within conflicting vehicle groups, deleting completely contained, covered, or highly overlapping areas to obtain updated conflict areas. This reduces the number of areas to be considered, lowers the scope of subsequent search calculations, avoids excessive computation, and makes the algorithm run more efficiently. When constructing the search tree, the vehicle passage order is determined based on the updated conflict areas, generating multiple candidate solutions, simplifying the search space, reducing the number of evaluation solutions, and improving execution efficiency. A preset priority passage strategy provides realistic passage criteria, ensuring priority passage for important or urgent vehicles. Combining the arrival time and passage time of vehicles in conflict areas allows for more accurate assessment of traffic conditions, ensuring that candidate solutions align with actual needs. A preset cost function is used to calculate the total cost of the complete passage sequence for each candidate solution, and the target solution is determined by minimizing the total cost. This evaluation method comprehensively considers factors such as vehicle waiting time, delays, and road resource utilization, optimizing overall traffic efficiency, reducing average vehicle waiting time and delays, and improving road capacity.

[0139] As an optional implementation, after determining the target cooperative scheduling scheme based on the above embodiments, the method further includes:

[0140] Extract the vehicles that need to give way and the conflict areas that need to give way from the target collaborative scheduling scheme;

[0141] Calculate the trajectory coordinates of the stopping point based on the entrance location of the conflict zone to be yielded to and the preset safety distance;

[0142] Project the trajectory coordinates of the stopping point onto the trajectory line to generate the stopping point;

[0143] If the conflict area satisfies the vehicle following relationship, no parking point will be generated.

[0144] Among them, vehicles that need to give way refer to those vehicles that are determined in the target coordinated scheduling scheme to need to wait for other high-priority vehicles to pass through the conflict area before they can continue to travel.

[0145] Among them, the conflict area that requires yielding refers to the specific conflict area in which a vehicle that needs to yield must wait for a vehicle with priority to pass during its journey, and it is the core basis for generating the stop point.

[0146] The stopping point refers to a temporary parking and waiting area planned for vehicles that need to give way. This location is situated at a safe distance in front of the entrance to the conflict zone to prevent vehicles from entering the conflict zone and causing traffic conflicts.

[0147] The preset safety distance refers to a fixed distance threshold that is set in advance, taking into account factors such as vehicle braking performance and reaction time.

[0148] Specifically, based on the established target collaborative scheduling scheme, all vehicles within the scheme that need to wait for other vehicles to pass first are selected, i.e., vehicles that need to yield. For each vehicle that needs to yield, the conflict areas that require waiting on its travel trajectory are further extracted. At the same time, these conflict areas are judged; if the two vehicles corresponding to the conflict area are in a following relationship, then the area is marked as not needing to generate a stopping point.

[0149] Furthermore, for conflict areas where there is no following relationship but waiting is necessary, the entry position parameters of the conflict area on the vehicle trajectory line are obtained. Using this entry position as a reference, the trajectory coordinates of candidate stopping points are calculated in conjunction with a preset safe backtracking distance. When a candidate stopping point is located within an intersection area, it can be further backtracked to a safe position outside the intersection or near the trajectory start point. Subsequently, these trajectory coordinates are projected onto the vehicle's planned trajectory line and converted into a two-dimensional stopping point in the world coordinate system. If the difference between the stopping point generated in the current cycle and the stopping point in the previous cycle is within a preset tolerance range, the stopping point from the previous cycle can be used to reduce command jitter.

[0150] The collaborative scheduling method for autonomous vehicles provided in this application calculates the trajectory coordinates of the stopping point based on the required yield point entrance location and a preset safety distance, and projects these coordinates onto the trajectory line to generate the stopping point. The safety distance setting ensures that the vehicle has sufficient braking distance before reaching the conflict area, enabling it to stop safely and preventing rear-end collisions and other accidents caused by insufficient stopping distance. The operations of extracting yield information, calculating stopping point coordinates, and generating the stopping point provide specific execution details for the target collaborative scheduling scheme. Based on these clear guidelines, the various requirements in the scheduling scheme can be accurately executed, improving the operability and executability of the scheduling scheme.

[0151] As an optional implementation, based on the above embodiments, it further includes:

[0152] Generate scheduling instructions based on the target-oriented collaborative scheduling scheme;

[0153] The dispatch instructions and stop point information are encapsulated into messages and sent to each vehicle through the communication link of the cloud dispatch service, so that each vehicle can drive according to the target collaborative dispatch plan.

[0154] Among them, the scheduling instruction refers to the set of instructions generated based on the target collaborative scheduling scheme and used to guide vehicle driving operations.

[0155] Specifically, the target coordinated scheduling scheme for each conflicting vehicle group is analyzed. For vehicles with priority right-of-way, a passage instruction is generated, which includes the conflict area identifier, the start time, and the expected passage duration. For vehicles that need to yield, a yield instruction is generated, which includes the conflict area identifier, the yield start time, and the associated information of the stopping point. Simultaneously, for vehicles in a following relationship, a following coordination instruction is generated, specifying parameters such as the target vehicle to follow and the safe following distance to avoid frequent starts and stops.

[0156] Furthermore, following a pre-defined standardized data format, the dispatch instructions for each vehicle are integrated and encapsulated with their corresponding stop point information. The cloud-based dispatch service is activated, establishing a communication link between the vehicle and the dispatch system. The encapsulated message is then sent to the corresponding planning and control module according to the vehicle's identifier. Upon receiving the message, the vehicle parses the instruction content, and the onboard control system adjusts its driving strategy accordingly. Priority vehicles maintain their planned speed, while yielding vehicles proceed to their stop points and wait until further instructions are received.

[0157] The collaborative scheduling method for autonomous vehicles provided in this application generates scheduling instructions based on a scientifically sound and reasonable target collaborative scheduling scheme. This provides precise driving guidance to each vehicle, specifying when, where, at what speed, and in what manner the vehicle will travel. Combined with stop point information, vehicles can prepare in advance, safely stopping or avoiding conflict zones, effectively preventing collisions and other safety accidents, and ensuring the safety of passengers and pedestrians. Real-time information sharing between vehicles is achieved through V2V message push. Vehicles can promptly understand the driving status and scheduling arrangements of other surrounding vehicles, detect potential hazards in advance, and further improve driving safety.

[0158] As an optional implementation, based on the above embodiments, it further includes:

[0159] Determine whether the conflict area calculated in the current cycle is consistent with that in the previous cycle;

[0160] Determine whether the vehicles and conflict relationships included in each conflicting vehicle group in the current cycle are consistent with those in the previous cycle;

[0161] Determine whether the target coordinated scheduling scheme of the previous period is in the set of candidate coordinated scheduling schemes for the current period;

[0162] Determine whether the difference between the total cost of the target coordinated scheduling scheme in the previous period and the total cost of the target coordinated scheduling scheme in the current period is less than a preset cost threshold;

[0163] If all the above conditions are met, the scheduling instruction of the previous cycle will be determined as the scheduling instruction of the current cycle.

[0164] The cycle refers to the time interval between the collaborative scheduling device of autonomous vehicles performing a complete scheduling calculation, which is the basic unit for updating scheduling instructions.

[0165] The set of candidate collaborative scheduling schemes refers to the set of all candidate scheduling schemes that satisfy the priority passage strategy generated by a pre-set greedy search algorithm within the current period.

[0166] Among them, the preset cost threshold refers to the pre-set critical value of cost difference, which is used to measure the degree of cost difference between the optimal scheduling schemes of two adjacent cycles.

[0167] Specifically, the core information, such as the spatial range and corresponding vehicle identification, of all conflict zones calculated in the current period is extracted. Then, the conflict zone data stored in the previous period is retrieved, and the conflict zones from the two periods are compared one by one to confirm whether the number, location range, and involved vehicles of the conflict zones are completely consistent. Only when all conflict zone information remains unchanged can the conflict zones be considered consistent. Based on the conflict zone data of the current period, the list of vehicles included in each conflict group is determined, as well as the direct and indirect conflict relationships among the vehicles within the group. This is then compared with the conflict group data of the previous period to verify whether the division results of the conflict groups, the vehicle members within the groups, and the conflict relationships between vehicles are completely matched. Only when all information is consistent can the conflict groups and conflict relationships be considered consistent.

[0168] Furthermore, after generating the set of candidate coordinated scheduling schemes for the current period, the target coordinated scheduling scheme for the previous period is extracted. The target scheme from the previous period is compared with all schemes in the set of candidate schemes for the current period to confirm whether there are any completely identical candidate schemes. The total cost of the target scheduling scheme from the previous period and the total cost of the target scheduling scheme for the current period are retrieved respectively. The absolute difference between the two total costs is calculated and compared with a preset cost threshold. If the difference is less than the preset cost threshold, it indicates that the cost difference between the optimal schemes of the two periods is within an acceptable range.

[0169] Understandably, the scheduling system will only trigger the scheduling solution preservation strategy if all four consistency conditions mentioned above are met simultaneously when executing the scheduling instruction reuse operation. In this case, the system will no longer issue new scheduling instructions generated in the current cycle, but will directly reuse the scheduling instructions from the previous cycle and send them to each vehicle. If any condition is not met, the system will issue new scheduling instructions for the current cycle normally, ensuring that vehicle driving strategies match real-time traffic conditions.

[0170] The collaborative scheduling method for autonomous vehicles provided in this application addresses traffic scenarios where, if the conflict area, vehicles in the conflicting vehicle groups, and conflict relationships are consistent between the current and previous cycles, and the target collaborative scheduling scheme from the previous cycle is in the candidate set with a total cost difference less than a preset threshold, it indicates that the traffic situation has not substantially changed. In this case, using the scheduling instructions from the previous cycle avoids recalculating the target collaborative scheduling scheme, reduces numerous complex calculations, significantly improves system computational efficiency, and reduces computational resource consumption. While ensuring the rationality of scheduling, it maximizes the continuity of scheduling instructions, significantly reduces the frequency of scheduling instruction switching, reduces frequent vehicle braking and starting / stopping, and improves travel comfort, energy efficiency, and road traffic efficiency.

[0171] Figure 3 A flowchart illustrating a collaborative scheduling method for autonomous vehicles provided in another embodiment of this application is shown below. Figure 3As shown, the cooperative scheduling method for autonomous vehicles provided in this embodiment includes the following steps:

[0172] S301. Obtain the preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters. The trajectory information includes the vehicle's positioning parameters and planned trajectory parameters.

[0173] S302. Based on the fitting of each trajectory point in the planned trajectory, generate the trajectory line and curvature curve corresponding to each vehicle.

[0174] S303. Calculate the vehicle bounding box corresponding to each trajectory point based on the vehicle configuration parameters, trajectory lines, and curvature curves.

[0175] S304. Combine the vehicle bounding boxes corresponding to each trajectory point to generate the trajectory geometric region corresponding to the vehicle. The trajectory geometric region is the spatial occupancy range of the vehicle within a preset time range.

[0176] S305. Calculate the spatial overlap of the geometric regions of the trajectories of any two vehicles.

[0177] S306. The following relationship between vehicles is determined based on the spatial overlap by adopting a preset following determination rule.

[0178] S307. Delete the spatial overlap that satisfies the vehicle following relationship, and project the boundary points of the remaining spatial overlap onto the corresponding trajectory line to obtain the conflict area.

[0179] S308. The vehicles are grouped according to the conflict area to obtain at least one conflict vehicle group, in which the vehicles in the conflict vehicle group have direct or indirect conflict relationships with each other.

[0180] S309. Calculate the expected speed of each trajectory point based on the vehicle's current positioning parameters, road speed limit, vehicle acceleration limit, and curvature curve.

[0181] S310. The average speed algorithm is used to calculate the time to reach each trajectory point based on the expected speed of each trajectory point, so as to obtain the time for the vehicle to reach each conflict area.

[0182] S311. Determine the time to pass through each conflict zone based on the arrival time of each trajectory point.

[0183] S312. Merge the conflict areas of each vehicle in the conflicting vehicle group, and delete conflict areas that completely contain, cover, or highly overlap to obtain updated conflict areas.

[0184] S313. Construct a search tree and determine the vehicle passage order in each updated conflict area according to the order of updating conflict areas to obtain multiple candidate collaborative scheduling schemes.

[0185] S314. Calculate the total cost corresponding to the complete passage sequence of each candidate cooperative scheduling scheme using a preset cost function.

[0186] S315. The candidate coordinated scheduling scheme with the lowest total cost is determined as the target coordinated scheduling scheme for the current conflicting train sets.

[0187] S316. Extract the vehicles that need to give way and the conflict areas that need to give way from the target coordinated scheduling scheme.

[0188] S317. Calculate the trajectory coordinates of the stopping point based on the entrance location of the conflict zone to be yielded to and the preset safety distance.

[0189] S318. Project the trajectory coordinates of the stopping point onto the trajectory line to generate the stopping point.

[0190] S319. Generate scheduling instructions based on the target collaborative scheduling scheme.

[0191] S320 encapsulates dispatch instructions and stop point information into messages and sends them to each vehicle through the communication link of the cloud dispatch service, so that each vehicle can drive according to the target collaborative dispatch plan.

[0192] In this embodiment, the implementation method and technical effect of S301-S320 are similar to those of the corresponding solutions in the above embodiments, and will not be repeated here.

[0193] Figure 4 This is a schematic diagram of the structure of the cooperative scheduling device for autonomous vehicles provided in this application, as shown below. Figure 4 As shown, the autonomous vehicle collaborative scheduling device 40 provided in this embodiment includes: an acquisition module 41, a calculation module 42, a grouping module 43, and a determination module 44.

[0194] The system includes: an acquisition module 41, which acquires a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters, including vehicle positioning parameters and planned trajectory parameters; a calculation module 42, which calculates the conflict area between any two vehicles based on the trajectory information and vehicle configuration parameters; a grouping module 43, which groups the vehicles according to the conflict area to obtain at least one conflict vehicle group, wherein the vehicles in the conflict vehicle group have direct or indirect conflict relationships with each other; the calculation module 42 is also used to calculate the arrival time of each vehicle in each conflict area and the duration of passage through the conflict area; and a determination module 44, which, for each conflict vehicle group, uses a preset partial greedy search algorithm based on the preset priority passage strategy, the arrival time of each vehicle in each conflict area, and the duration of passage through the conflict area to determine a target cooperative scheduling scheme so that each vehicle travels according to the target cooperative scheduling scheme.

[0195] The collaborative scheduling device for autonomous vehicles provided in this embodiment can perform... Figure 2 The implementation principles and technical effects of the methods shown are similar, and will not be repeated here.

[0196] Optionally, the vehicle configuration parameters include vehicle length, width, and wheelbase. The calculation module 42, when calculating the conflict area between any two vehicles based on trajectory information and vehicle configuration parameters, specifically performs the following: generating trajectory lines and curvature curves corresponding to each vehicle based on the fitting of trajectory points in the planned trajectory; calculating the vehicle bounding box corresponding to each trajectory point based on the vehicle configuration parameters, trajectory lines, and curvature curves; combining the vehicle bounding boxes corresponding to each trajectory point to generate the trajectory geometric region corresponding to the vehicle, where the trajectory geometric region is the spatial occupancy range of the vehicle within a preset time range; calculating the spatial overlap of the trajectory geometric regions of any two vehicles; determining the vehicle following relationship based on the spatial overlap using preset following judgment rules; deleting the spatial overlap that satisfies the vehicle following relationship, and projecting the boundary points of the remaining spatial overlap onto the corresponding trajectory lines to obtain the conflict area.

[0197] Optionally, the calculation module 42, when calculating the time for each vehicle to arrive at each conflict zone and the duration of passage through the conflict zone, specifically performs the following: calculates the expected speed of each trajectory point based on the vehicle's current positioning parameters, road speed limit, vehicle acceleration limit, and curvature curve; calculates the time to arrive at each trajectory point based on the expected speed of each trajectory point using an average speed algorithm, so as to obtain the time for the vehicle to arrive at each conflict zone; and determines the duration of passage through each conflict zone based on the time to arrive at each trajectory point.

[0198] Optionally, the determining module 44, when using a preset partial greedy search algorithm to determine the target coordinated scheduling scheme based on a preset priority passage strategy, the arrival time of each vehicle in each conflict area, and the duration of passage through the conflict area, specifically performs the following: merging the conflict areas of each vehicle in the conflicting vehicle group, deleting conflict areas that completely contain, cover, or highly overlap, to obtain updated conflict areas; constructing a search tree and determining the vehicle passage order in each updated conflict area according to the order of updating the conflict areas, to obtain multiple candidate coordinated scheduling schemes; calculating the total cost corresponding to the complete passage sequence of each candidate coordinated scheduling scheme using a preset cost function; and determining the candidate coordinated scheduling scheme with the minimum total cost as the target coordinated scheduling scheme for the current conflicting vehicle group.

[0199] Optionally, the collaborative scheduling device for autonomous vehicles provided in this embodiment further includes an extraction module and a generation module.

[0200] Correspondingly, the extraction module is used to extract the vehicles that need to give way and the conflict areas that need to give way from the target collaborative scheduling scheme; the calculation module 42 is also used to calculate the trajectory coordinates of the stopping point based on the entrance position of the conflict area that needs to give way and the preset safety distance; the generation module is used to project the trajectory coordinates of the stopping point onto the trajectory line to generate the stopping point; if it is a conflict area that satisfies the vehicle following relationship, no stopping point is generated.

[0201] Optionally, the collaborative scheduling device for autonomous vehicles provided in this embodiment further includes a sending module.

[0202] Correspondingly, the generation module is used to generate scheduling instructions based on the target collaborative scheduling scheme; the sending module is used to encapsulate the scheduling instructions and stop point information into messages and send them to each vehicle through the communication link of the cloud scheduling service so that each vehicle can drive according to the target collaborative scheduling scheme.

[0203] Optionally, the determining module 44 is further configured to determine whether the conflict area calculated in the current cycle is consistent with that in the previous cycle; determine whether the vehicles and conflict relationships included in each conflicting train group in the current cycle are consistent with those in the previous cycle; determine whether the target coordinated scheduling scheme of the previous cycle is in the set of candidate coordinated scheduling schemes in the current cycle; determine whether the difference between the total cost of the target coordinated scheduling scheme of the previous cycle and the total cost of the target coordinated scheduling scheme of the current cycle is less than a preset cost threshold; if all the above conditions are met, then the scheduling instruction of the previous cycle is determined as the scheduling instruction of the current cycle.

[0204] Figure 5 This is a schematic diagram of the structure of the collaborative scheduling device for autonomous vehicles provided in this application. Figure 5 As shown, the collaborative scheduling device 50 for autonomous vehicles provided in this embodiment includes a processor 51 and a memory 52. ​​The processor 51 and the memory 52 are connected via a bus and communicate with each other.

[0205] In the specific implementation process, the processor 51 executes the computer execution instructions stored in the memory 52, causing the processor 51 to perform the above-described method.

[0206] The specific implementation process of processor 51 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0207] In the above embodiments, it should be understood that the processor 51 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0208] The memory 52 may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0209] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0210] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0211] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0212] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0213] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0214] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0215] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0216] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0217] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0218] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0219] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for cooperative scheduling of autonomous vehicles, characterized in that, The method includes: The system acquires a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters, wherein the trajectory information includes vehicle positioning parameters and planned trajectory parameters. The conflict zone between any two vehicles is calculated based on the trajectory information and vehicle configuration parameters. The vehicles are grouped according to the conflict area to obtain at least one conflict vehicle group, in which the vehicles in the conflict vehicle group have direct or indirect conflict relationships with each other. Calculate the time it takes for each vehicle to arrive at each conflict zone and the duration it takes to pass through that conflict zone; For each conflicting vehicle group, a pre-set partial greedy search algorithm is used to determine the target coordinated scheduling scheme based on the pre-set priority passage strategy, the arrival time of each vehicle in each conflict area, and the duration of passage through the conflict area, so that each vehicle travels according to the target coordinated scheduling scheme.

2. The method according to claim 1, characterized in that, The vehicle configuration parameters include vehicle length, width, and track width; the calculation of the conflict area between any two vehicles based on the trajectory information and vehicle configuration parameters includes: The trajectory lines and curvature curves corresponding to each vehicle are generated by fitting each trajectory point in the planned trajectory. Calculate the vehicle bounding box corresponding to each trajectory point based on the vehicle configuration parameters, trajectory lines, and curvature curves; The vehicle bounding boxes corresponding to each trajectory point are combined to generate the trajectory geometric region corresponding to the vehicle. The trajectory geometric region is the spatial occupancy range of the vehicle within a preset time range. Calculate the spatial overlap of the geometric regions of the trajectories of any two vehicles; The vehicle following relationship is determined based on the spatial overlap using a preset following determination rule; The overlapping spatial portions that satisfy the vehicle following relationship are deleted, and the boundary points of the remaining overlapping spatial portions are projected onto the corresponding trajectory lines to obtain the conflict area.

3. The method according to claim 2, characterized in that, The calculation of the arrival time of each vehicle in each conflict zone and the duration of passage through that conflict zone includes: Calculate the expected speed at each trajectory point based on the vehicle's current positioning parameters, road speed limit, vehicle acceleration limit, and curvature curve; The average speed algorithm is used to calculate the time to reach each trajectory point based on the expected speed of each trajectory point, so as to obtain the time for the vehicle to reach each conflict area. The time required to pass through each conflict zone is determined based on the arrival time at each trajectory point.

4. The method according to claim 1, characterized in that, The method employs a pre-defined partially greedy search algorithm to determine a target cooperative scheduling scheme based on the pre-defined priority passage strategy, the arrival time of each vehicle in each conflict zone, and the duration of passage through the conflict zone. This includes: The conflict areas of each vehicle in the conflict group are merged, and conflict areas that completely contain, cover, or highly overlap are deleted to obtain updated conflict areas. Construct a search tree and determine the vehicle passage order in each updated conflict area according to the order of updating the conflict areas to obtain multiple candidate cooperative scheduling schemes; The total cost corresponding to the complete passage sequence of each candidate cooperative scheduling scheme is calculated using a preset cost function; The candidate coordinated scheduling scheme with the lowest total cost is determined as the target coordinated scheduling scheme for the current conflicting train crews.

5. The method according to claim 2, characterized in that, After determining the target collaborative scheduling scheme, the method further includes: Extract the vehicles that need to give way and the conflict areas that need to give way from the target collaborative scheduling scheme; Calculate the trajectory coordinates of the stopping point based on the entrance location of the conflict zone that needs to be yielded to and the preset safety distance; The trajectory coordinates of the stopping point are projected onto the trajectory line to generate the stopping point; If the conflict area satisfies the vehicle following relationship, no parking point will be generated.

6. The method according to claim 5, characterized in that, Also includes: Generate scheduling instructions based on the target collaborative scheduling scheme; The dispatch instructions and stop point information are encapsulated into messages and sent to each vehicle through the communication link of the cloud dispatch service, so that each vehicle can drive according to the target collaborative dispatch plan.

7. The method according to claim 6, characterized in that, Also includes: Determine whether the conflict area calculated in the current cycle is consistent with that in the previous cycle; Determine whether the vehicles and conflict relationships included in each conflicting vehicle group in the current cycle are consistent with those in the previous cycle; Determine whether the target coordinated scheduling scheme of the previous period is in the set of candidate coordinated scheduling schemes for the current period; Determine whether the difference between the total cost of the target coordinated scheduling scheme in the previous period and the total cost of the target coordinated scheduling scheme in the current period is less than a preset cost threshold. If all the above conditions are met, the scheduling instruction of the previous period will be determined as the scheduling instruction of the current period.

8. A collaborative scheduling device for autonomous vehicles, characterized in that, include: The acquisition module is used to acquire a preset priority passage strategy, trajectory information of multiple vehicles, and vehicle configuration parameters. The trajectory information includes vehicle positioning parameters and planned trajectory parameters. The calculation module is used to calculate the conflict area between any two vehicles based on the trajectory information and vehicle configuration parameters. The grouping module is used to group the vehicles according to the conflict area to obtain at least one conflict vehicle group, wherein the vehicles in the conflict vehicle group have a direct or indirect conflict relationship with each other. The calculation module is also used to calculate the time it takes for each vehicle to arrive at each conflict zone and the duration of passing through the conflict zone. The determination module is used to determine the target coordinated scheduling scheme for each conflicting vehicle group by using a preset partial greedy search algorithm based on the preset priority passage strategy, the time of each vehicle to arrive at each conflict area and the duration of passing through the conflict area, so that each vehicle travels according to the target coordinated scheduling scheme.

9. A collaborative scheduling device for autonomous vehicles, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

11. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-7.

Images