Vehicle transportation task seamless scheduling control method based on space-time three-dimensional collaborative optimization

Abstract

The present application relates to a kind of seamless scheduling control method of vehicle transportation task based on space-time stereoscopic coordination optimization, comprising: constructing task operation point set and space-time constraint set;Under the space-time constraint set, with minimizing vehicle completion time and task gap time as optimization goal, the task sequence corresponding to each operation point is solved;For each operation point corresponding task sequence, time scheduling is carried out, for the adjacent task corresponding to the same operation point, seamless connection verification is carried out, and when verification fails, local scheme adjustment is carried out, based on all the task sequence of successful verification, the working state of each vehicle is controlled accordingly, and the closed-loop control is implemented to the driving process of vehicle between operation node.Compared with prior art, the present application can realize the space-time accurate matching of transportation vehicle and transportation task by the space-time accurate control of automatic driving transport vehicle, effectively avoid task discontinuity and vehicle waiting, so that task is continuous and executable on time axis and realizes seamless connection.

Description

Technical Field

[0001] This invention relates to the field of road transport vehicle driving control technology, and in particular to a seamless scheduling and control method for vehicle transport tasks based on spatiotemporal three-dimensional collaborative optimization. Background Technology

[0002] The management and control of transport vehicles is a key technology in modern logistics and production operation systems, widely applied in scenarios such as port collection and distribution and yard transshipment, short-haul transportation between industrial parks and mining areas, urban distribution and trunk-branch line connections, and warehousing handling and loading / unloading coordination. Its core function is to achieve efficient utilization of transportation resources and timely delivery of transportation services by rationally controlling the routes, sequences, and timing of vehicle task execution, given limited vehicle resources and constraints on task arrival and operating conditions. This reduces operating costs, minimizes empty runs and waiting times, and improves system throughput and turnover efficiency.

[0003] In traditional transportation systems primarily reliant on manually driven vehicles, vehicle movement and arrival are significantly uncertain due to factors such as driver habits, speed selection, route preferences, and random on-site disturbances. Particularly at task handover points, the arrival time, parking location, and pick-up actions are difficult to maintain consistent, making it challenging for management to impose strictly enforceable constraints on task start / end times. Given these uncertainties, existing technologies in vehicle transportation management and control typically focus on vehicle-cargo matching / task allocation and planning of travel routes or work sequences—that is, determining "which vehicle performs which tasks, in what order, and along which route." However, such planning often fails to guarantee continuous temporal connections between tasks: because vehicle arrival and pick-up times are difficult to precisely constrain, waiting and gaps easily occur between tasks, preventing adjacent tasks from forming a stable, continuous execution chain, thus reducing the implementability and efficiency of management solutions.

[0004] With the development of autonomous driving technology, autonomous transport vehicles are now being used in relatively closed or semi-closed environments such as ports, mining areas, and industrial parks to perform related transportation tasks. This means that transport vehicles are gradually evolving from traditional "manually driven vehicles executing control commands" to "executors that can be precisely controlled by a control system." Autonomous transport vehicles can perform closed-loop adjustments to their speed and arrival behavior under the constraints of the control system, thereby making the arrival time and pick-up actions at work nodes more controllable and repeatable. This will help reduce waiting and downtime, improve task chain continuity, and enhance system turnaround efficiency.

[0005] Although autonomous transport vehicles have the characteristic of precise control, ensuring the feasibility of vehicle-cargo matching and task sequence in actual operation scenarios, while achieving timely arrival and pickup of vehicles at work nodes, and enabling continuous connection of adjacent tasks at work nodes, remains a technical challenge that urgently needs to be solved. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology by providing a seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization. This method can accurately match autonomous driving vehicles with transportation tasks in time and space and synchronize task planning, effectively avoiding task interruptions and vehicle waiting, and enabling tasks to be continuously executable and seamlessly connected on the timeline.

[0007] The objective of this invention can be achieved through the following technical solution: a seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization, comprising the following steps:

[0008] S1. Obtain the vehicle set and task set, and combine them with the spatiotemporal constraints to construct the task operation point set and the spatiotemporal constraint set;

[0009] S2. Under the set of spatiotemporal constraints, with the optimization objective of minimizing vehicle completion time and minimizing task gap time, the task sequence corresponding to each work point is generated by solving the problem.

[0010] S3. Arrange the time sequence for the task sequence corresponding to each work point, calculate the start and finish times of each task in the task sequence, so that adjacent tasks at the work point are connected continuously on the time axis.

[0011] S4. Perform seamless connection verification on adjacent tasks at the same work point to determine whether the task sequence at the work point is continuous and executable. If the verification fails, proceed to step S5. If the verification succeeds, proceed to step S6.

[0012] S5. Adjust the local scheme for the task sequence corresponding to the job point that failed the verification, and then return to step S3;

[0013] S6. Based on the task sequence corresponding to all successfully verified work points, control the working status of each vehicle accordingly, and implement closed-loop control of the vehicle's driving process between work points.

[0014] Furthermore, the vehicle set in S1 specifically refers to... It includes several autonomous transport vehicles, each of which It has the following attributes: vehicle initial position Available start time and from any position Arrive at the location travel time ;

[0015] The task set is specifically as follows It contains several tasks, any one of which It has the following attributes: starting point ,end Task duration and time window The time window represents the range of times during which a task can begin.

[0016] Furthermore, the set of task operation points constructed in S1 is as follows: This includes multiple work points where vehicles pick up tasks;

[0017] The spatiotemporal constraint set constructed in S1 is as follows:

[0018]

[0019] in, This indicates the task-vehicle matching constraint. This indicates a sequence constraint of work points. Indicates time window constraints. Represents spatial reachability constraints;

[0020] To ensure that each task at each work point is assigned to one vehicle, and that the same vehicle performs only one task at any given time, it is specifically represented as follows: In the formula, Assign variables to the task, when the vehicle At the work site Execute the task hour ,otherwise ;

[0021] This is used to ensure that each task is assigned to a corresponding job point, and that the task order at each job point must satisfy the timing constraint, specifically expressed as: , , , In the formula, where, Assign variables to tasks at work points if and only if the task Assigned to work site hour ; For task adjacency variables, if and only if at the job point Location, Task After being assigned, the task immediately followed. If also assigned, then ; As the first task execution variable, when the task... For work points When the first task is assigned, The relationship between the above variables is as follows: , ;

[0022] This is used to ensure that the start time of each task must satisfy the time window constraint, specifically expressed as: In the formula, Indicates task The beginning moment;

[0023] This is used to ensure that vehicles must meet spatial accessibility constraints while performing tasks at work sites, and it is specifically expressed as follows: In the formula, Indicates task The completion time.

[0024] Furthermore, the optimization objective in S2 is specifically:

[0025]

[0026]

[0027]

[0028]

[0029] in, This indicates the completion time of the last task. Indicates work point Task With the task The gap time between them;

[0030] Specifically, S2 employs a genetic algorithm to solve the optimization objective under a set of spatiotemporal constraints, thereby generating a vehicle task matching sequence corresponding to each work point.

[0031] Furthermore, the process of solving the optimization objective using a genetic algorithm under the set of spatiotemporal constraints in S2 includes:

[0032] First, an initial population is generated. Each chromosome in the population represents a vehicle-cargo matching scheme and task sequence arrangement. Each chromosome consists of two parts, including a representation of the job point. Task j Assigned to vehicles i Task assignment coding And indicates the same work point p Task sequence coding of adjacent connection relationships at the location ;

[0033] The fitness of each chromosome in the initial population is then evaluated, and the fitness function is designed according to the optimization objective:

[0034]

[0035] Chromosomes with higher fitness indicate less vehicle completion and waiting time. Based on the fitness function value, selection, crossover, and mutation operations are performed to generate a new generation of the population. This new generation will continue to undergo fitness evaluation, selection, crossover, and mutation until the optimal solution that satisfies the spatiotemporal constraints and achieves the target fitness value is found. The final vehicle task matching sequence at each task operation point is represented as a set: For any work point : , It is a set, and the job points are given at the same time. p Relevant tasks j By vehicle i Acceptance, and adjacent tasks j and k The order in which they occur.

[0036] When generating a new generation of individuals, the selection operation refers to selecting individuals with higher fitness to enter the next generation based on the value of the fitness function; the crossover operation refers to combining the task allocation and task order information of two parent individuals to generate new offspring individuals. For example, in the task allocation part, a new task allocation scheme is generated by exchanging the task allocation codes in the parent individuals, and in the task order part, a new task execution order is generated by exchanging the task order codes in the parent individuals; the mutation operation generates new individuals by randomly changing the allocation or order of certain tasks. For example, randomly selecting a task and changing the vehicle it is assigned to, or randomly selecting a vehicle and changing its task order.

[0037] Furthermore, the specific process of S3 is as follows:

[0038] First, calculate the start time of the first task for each work point: ,in, For work points The moment the first mission begins, For vehicles Available start time, It is a vehicle From its initial position To the work site Travel time required to reach the location;

[0039] The subsequent calculation of the start time of the task is as follows: ,in, For work points Complete the task Next task At the beginning of Indicates task The time of receipt, Indicates that the vehicle is on the mission Location To the mission Location Required travel time;

[0040] Through the above process, the start and end times of each task at each work point are obtained. Based on the task sequence obtained in S2, the corresponding time sequence is assigned to the tasks at each work point, resulting in a task chain that satisfies the consistency of time sequence, ensuring that the tasks at each work point are connected in time sequence.

[0041] Furthermore, the seamless connection verification of adjacent tasks at the same work point in S4 includes work point continuity verification, time window consistency verification, and vehicle arrival point reachability verification. Specifically, the work point continuity verification is for adjacent task pairs at the same work point, requiring that the work gap between the planned start time of the subsequent task and the completion time of the preceding task does not exceed a preset threshold, and avoiding overlapping work intervals, so as to ensure continuous operation at work points without resource conflicts.

[0042] The time window consistency check specifically requires that for any task, the time window constraint in the spatiotemporal constraint set must be satisfied to ensure that the start time of the plan falls within the task's allowed time window.

[0043] The vehicle reachability verification specifically targets each vehicle undertaking adjacent tasks. It must satisfy the work point sequence constraint and spatial reachability constraint to ensure that the vehicle can arrive at the work point and complete the task acceptance before the start time of its corresponding task, so that the work point continuity requirement can be actually executed.

[0044] Furthermore, the local scheme adjustment in S5 includes sequence adjustment, reallocation, and time fine-tuning processes.

[0045] Furthermore, the sequence adjustment specifically involves, under the premise of a set of spatiotemporal constraints, locally exchanging, inserting, or reconnecting the task order at the same work point, so that the adjacent relationships that cause the work gap to exceed the threshold or the work interval to conflict are replaced with adjacent relationships that meet the requirements of continuity and non-conflict.

[0046] The reassignment specifically refers to tasks that violate time windows or reachability constraints. Adjust its assignment variables And update the associated ones simultaneously. The variable allows the task to be transferred to a vehicle that can arrive at the corresponding work point at the scheduled time, in order to fill the work gap and restore continuous operation;

[0047] Specifically, the time fine-tuning involves locally compressing or shifting the start / end times of related tasks at the same work point without violating the time window constraint, so that the work point continuity requirement and the timing consistency constraint are met simultaneously.

[0048] Furthermore, the specific process of S6 includes:

[0049] For vehicles By task Move to the next task During the driving process, the task is set. The completion time is ,Task The start time is , Indicates the length of time the vehicle is used for transfer, within a time interval. Based on the vehicle's real-time path progress, position, speed, and acceleration, the system continuously calculates and applies control parameters to ensure the vehicle stays on track at all times. Meet the on-time access conditions and implement closed-loop control as follows:

[0050] Assume the vehicle starts from the mission End node to task The planned driving path length at the starting node is Introduce progress coordinates along this path ,in Indicates that it is located in the task End node, Indicates arrival at the mission Starting node; vehicle At any moment The progress position is denoted as Speed ​​is denoted as Acceleration is denoted as ;

[0051] To achieve "on-time arrival", a reference schedule with time calibration is constructed. Define reference speed Schedule error With speed error The autonomous driving controller of the autonomous transport vehicle calculates and controls the acceleration in real time based on the error in each control cycle:

[0052]

[0053] in , To control the gain, Given an acceleration limiting function, the vehicle motion satisfies:

[0054]

[0055] When the arrival time If the on-time pickup condition is met:

[0056]

[0057] Then determine the vehicle In the mission The starting node arrives on time and meets the acceptance conditions, triggering the task. The task is received and the next task control process begins; among which... To allow for tolerance at the destination location, For speed tolerance, The target speed allowed when accepting the task;

[0058] If the conditions for acceptance are not met, it is determined that there is a deviation in the current "task chain and timestamp" at the execution layer, and the rollback process is entered. The arrival point deviation and speed deviation of the vehicle are calculated, the minimum time margin that needs to be compensated is calculated, and the minimum time margin is fed back to the spatiotemporal constraint set to update the subsequent constraints, and then the process returns to step S2.

[0059] The formula for calculating the minimum time margin that needs to be compensated is as follows: .

[0060] Compared with the prior art, the present invention has the following advantages:

[0061] This invention first constructs a set of task operation points and a set of spatiotemporal constraints. Then, under the set of spatiotemporal constraints, it generates a task sequence corresponding to each operation point by minimizing vehicle completion time and minimizing task gap time as optimization objectives. Next, it performs time-based orchestration on the task sequences corresponding to each operation point and performs seamless connection verification on adjacent tasks at the same operation point. Finally, based on the task sequences corresponding to all successfully verified operation points, it controls the working state of each vehicle accordingly, implementing closed-loop control of the vehicle's travel process between operation nodes. This achieves an integrated solution of "precise spatiotemporal matching of vehicles and goods and time-level task orchestration," accurately matching autonomous driving transport vehicles and transport tasks in the spatiotemporal dimension, and enabling timely arrival and pickup of vehicles at operation nodes. This allows for "seamless connection" between adjacent tasks at nodes, effectively avoiding task interruptions and vehicle waiting, and improving the continuity of transport task sequences.

[0062] After acquiring the vehicle set and the task set, this invention organizes information such as task start and end locations, task time windows, task duration, road network travel time, and vehicle availability. Based on this, it constructs a task operation point set and a unified spatiotemporal constraint set. The spatiotemporal constraint set includes task-vehicle matching constraints, task sequence constraints (i.e., time sequence constraints), time window constraints, and spatial reachability constraints. This systematically eliminates potential waiting, conflict, and chain break risks in the task chain of operation points, ensuring that tasks are continuously executable on the time axis and achieving seamless connection and stable serialization of vehicle-cargo matching tasks.

[0063] This invention takes minimizing vehicle completion time and minimizing task gap time as optimization objectives. It combines spatiotemporal constraints to solve for the vehicle task matching sequence at each task operation point. Then, it performs time-based arrangement of the task sequence at each task operation point. By calculating the task start time and completion time at each operation point and combining the task sequence, it assigns the corresponding time sequence to the task at each operation point, and finally obtains a task chain that satisfies time sequence consistency, which can ensure that the tasks at each operation point are connected in time sequence.

[0064] This invention performs seamless connection verification on task sequences undertaken by multiple vehicles at the same task operation point, including operation point continuity verification, time window consistency verification, and vehicle arrival point reachability verification, which can reliably ensure the continuous execution of the task chain at the operation point on the time axis.

[0065] This invention implements closed-loop control on the vehicle's travel process between adjacent work nodes based on the task sequence corresponding to all successfully verified work points. Specifically, it takes "the next task must arrive at its starting node and meet the pick-up conditions" as the control objective, constructs a reference arrival process based on available time and travel distance between nodes, and performs closed-loop correction during vehicle travel. Through closed-loop control, the autonomous transport vehicle can accurately arrive at the work node at the corresponding task start time and pick up the task, achieving seamless connection between adjacent tasks. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0067] Figure 2 This is a schematic diagram of the application framework for an example. Detailed Implementation

[0068] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0069] Example

[0070] This invention addresses the core challenge of driving control for road transport vehicles. It upgrades existing solutions that only output "vehicle-cargo matching results / task allocation results" to an integrated solution of "precise spatiotemporal matching of vehicles and cargo and time-level task orchestration." It proposes a seamless orchestration control method for vehicle transport tasks based on spatiotemporal three-dimensional collaborative optimization, such as... Figure 1 As shown, it includes the following steps:

[0071] S1. Obtain the vehicle set and task set, and combine them with the spatiotemporal constraints to construct the task operation point set and the spatiotemporal constraint set;

[0072] S2. Under the set of spatiotemporal constraints, with the optimization objective of minimizing vehicle completion time and minimizing task gap time, the task sequence corresponding to each work point is generated by solving the problem.

[0073] S3. Arrange the time sequence for the task sequence corresponding to each work point, calculate the start and finish times of each task in the task sequence, so that adjacent tasks at the work point are connected continuously on the time axis.

[0074] S4. Perform seamless connection verification on adjacent tasks at the same work point to determine whether the task sequence at the work point is continuous and executable. If the verification fails, proceed to step S5. If the verification succeeds, proceed to step S6.

[0075] S5. Adjust the local scheme for the task sequence corresponding to the job point that failed the verification, and then return to step S3;

[0076] S6. Based on the task sequence corresponding to all successfully verified work points, control the working status of each vehicle accordingly, and implement closed-loop control of the vehicle's driving process between work points.

[0077] The core of the above scheme lies in not only planning the matching relationship between vehicles and goods / transportation tasks, but also simultaneously planning the specific time when vehicles execute each transportation task, so that the vehicle-cargo matching tasks are seamlessly connected on the timeline. Given a set of autonomous driving transport vehicles, a set of goods / transportation tasks, and spatiotemporal constraints, this scheme can simultaneously complete the spatiotemporal matching of vehicles and goods and the planning of specific task times, output a timestamped task chain for each work point, and perform connection feasibility verification on adjacent tasks at each work point, so that the task chain is continuous and executable.

[0078] This embodiment applies the above-described solution, such as Figure 2 As shown, the main processes include:

[0079] First, obtain the vehicle set and task set, organize information such as task start and end locations, task time windows, task duration, road network travel time and vehicle availability, and on this basis, construct a task operation point set and a unified spatiotemporal constraint set.

[0080] First, define the relevant information regarding the task set and the set of autonomous transport vehicles: Task set It contains several tasks, any one of which It has the following attributes: starting point ,end Task duration and time window This indicates the time range within which the task can begin;

[0081] vehicle assembly It includes several autonomous transport vehicles, each of which It has the following attributes: vehicle initial position Available start time and from any position Arrive at the location travel time .

[0082] Next, construct the task assignment point set: This indicates the location (i.e., the work site) where the vehicle picks up the task.

[0083] Based on the above information, the spatiotemporal constraint set is constructed as follows:

[0084] (1)

[0085] in, This indicates the task-vehicle matching constraint. This indicates a sequence constraint of work points. Indicates time window constraints. This represents spatial reachability constraints.

[0086] To ensure that each task at each work point is assigned to one vehicle, and that the same vehicle performs only one task at any given time, it is specifically represented as follows:

[0087] (2)

[0088] in, Assign variables to the task, when the vehicle Execute the task hour ,otherwise .

[0089] This is used to ensure that each task is assigned to a corresponding job point, and that the task order at each job point must satisfy the timing constraint, specifically expressed as:

[0090] (3)

[0091] (4)

[0092] (5)

[0093] (6)

[0094] Assign variables to tasks at work points if and only if the task Assigned to work site hour ; For task adjacency variables, if and only if at the job point Location, Task After being assigned, the task immediately followed. If also assigned, then ; As the first task execution variable, when the task... For work points When the first task is assigned, The relationship between the above variables is as follows: , ; This is used to ensure that the start time of each task must satisfy the time window constraint, specifically expressed as:

[0095] (7)

[0096] in Indicates task The beginning of the moment.

[0097] This is used to ensure that vehicles must meet spatial accessibility constraints while performing tasks at work sites, and it is specifically expressed as follows:

[0098] (8)

[0099] in Indicates task The completion time.

[0100] This scheme explicitly models and verifies "seamless connection" as the core objective and constraint system for vehicle task orchestration. For the feasibility of connecting adjacent tasks at each work point, joint constraints are applied from three dimensions: spatial accessibility, time window consistency, and resource usage non-conflict. Equations (3) to (6) ensure the temporal constraints that the task sequence at each work point must meet; equation (7) guarantees that the task start time meets the time window constraint; and equation (8) ensures that each vehicle must meet the spatial accessibility constraint during task execution. This systematically eliminates potential waiting, conflict, and chain break risks in the task chain at work points, enabling tasks to be continuously executable on the timeline and achieving seamless connection and stable serialization of vehicle-cargo matching tasks.

[0101] Second, under the constructed set of spatiotemporal constraints, the task allocation and connection order are jointly optimized to generate the task sequence at each task operation point.

[0102] This solution takes "minimizing vehicle completion time" and "minimizing task idle time" as optimization objectives:

[0103] Minimize vehicle completion time: (9)

[0104] Minimize task gap time: (10)

[0105] By minimizing To ensure that vehicles can quickly receive and execute tasks at work sites, including Indicates the completion time of the last task: And by minimizing To reduce the time gap between tasks, among which Indicates work point Task With the task The gap time between: .

[0106] This embodiment uses a genetic algorithm on a set of spatiotemporal constraints. The optimization objective described above is solved as follows:

[0107] First, an initial population is generated, containing multiple individuals, i.e., multiple chromosomes. Each chromosome represents a vehicle-cargo matching scheme and task order arrangement. Each chromosome consists of two parts, including a task allocation code indicating which vehicle the task is assigned to. And task sequence coding that represents the adjacent connection relationship at the same work point In this embodiment, an initial population is generated randomly, with each individual representing a possible vehicle-cargo matching relationship and task sequence arrangement. Subsequently, the fitness of each individual in the initial population is evaluated, and the fitness function is designed according to the optimization objective.

[0108] (11)

[0109] Individuals with higher fitness indicate less vehicle completion and waiting time. Then, based on the fitness function value, selection, crossover, and mutation operations are performed to generate a new generation of the population. This new generation will continue to undergo fitness evaluation, selection, crossover, and mutation until the optimal solution that satisfies the spatiotemporal constraints and achieves the target fitness value is found.

[0110] The final generated vehicle task matching sequence at each task operation point is represented as a set: For any work point :

[0111] (12)

[0112] It is a set that simultaneously indicates which vehicle will undertake the relevant tasks at that work point, as well as the order in which adjacent tasks will be performed.

[0113] Third, the task sequence at each generated task work point is time-arranged, and the start and end times of related tasks are calculated so that, under the premise of satisfying relevant constraints, adjacent tasks at the work point are connected continuously on the time axis.

[0114] First, calculate the start time of the first task for each work point:

[0115] (13)

[0116] in, For work points The moment the first mission begins, For vehicles Available start time, It is a vehicle From its initial position To the work site The travel time required to reach the location.

[0117] Then, the start time of subsequent tasks is calculated, specifically as follows:

[0118] (14)

[0119] in, For work points Complete the task Next task At the beginning of Indicates task The time of receipt, Indicates that the vehicle is on the mission Location To the mission Location Required travel time.

[0120] Through the above process, the start and end times of each task at each work point are obtained. Based on the obtained task sequence, a corresponding time sequence is assigned to the tasks at each work point, thus forming a task chain that satisfies the consistency of time sequence and ensures that the tasks at each work point are connected in time sequence.

[0121] This solution utilizes task assignment variables. Adjacency variables of tasks A task succession chain structure is formed for each work point, and the start and end times of the tasks are explicitly calculated using equations (13) and (14), thereby outputting an executable scheduling plan for the "task chain and timestamp" at each work point. This avoids the problem of unimplementability caused by only giving the order but lacking specific times, and makes the vehicle-cargo matching decision directly have time feasibility and continuity at the execution level.

[0122] Fourth, perform seamless connection verification on adjacent tasks at the same task work point to determine whether the task chain at the work point is continuous and executable.

[0123] It should be noted that the task operation point is the loading / unloading point or service point corresponding to the task start / end point, and adjacent tasks refer to two tasks that are adjacent to each other at the same operation point according to the planned operation sequence.

[0124] The seamless integration verification in this solution includes at least the following:

[0125] Job continuity verification: For adjacent task pairs at the same job point, the job gap between the planned start time of the subsequent task and the completion time of the preceding task should not exceed a preset threshold, and job intervals should be avoided to ensure continuous operation at job points without resource conflicts.

[0126] Time window consistency check: For any task, the time window constraint (7) in the spatiotemporal constraint set must still be satisfied to ensure that the start time of the plan falls within the task's allowed time window;

[0127] Vehicle reachability verification: For each vehicle that undertakes an adjacent task, it is still necessary to satisfy the temporal constraints (3) to (6) and spatial reachability constraints (8) corresponding to the given task order and time arrangement results, so as to ensure that the vehicle can arrive at the work point and complete the task acceptance before the start time of its corresponding task, so that the work point continuity requirement can be actually executed.

[0128] 5. When the seamless connection verification fails, the local solution is adjusted to ensure that adjacent tasks at the same work point meet the condition of continuous execution.

[0129] The adjustments made to the partial plan include:

[0130] Sequence adjustment: Under the premise that the unique assignment constraint (1) of the task is maintained, the task operation sequence at the same operation point is locally exchanged, inserted or reconnected, so that the adjacent relationship that causes the operation gap to exceed the threshold or the operation interval to conflict is replaced with the adjacent relationship that meets the requirements of continuity and non-conflict.

[0131] Reallocation: For tasks that violate time windows or reachability constraints. Adjust its assignment variables And update the associated ones simultaneously. The variable allows the task to be transferred to a vehicle that can arrive at the corresponding work point at the scheduled time, in order to fill the work gap and restore continuous operation;

[0132] Time fine-tuning: Without violating constraint (7), the start / end time of related tasks at the same work point is locally compressed or shifted so that the work point continuity requirement and the time sequence consistency constraint are met at the same time.

[0133] Then, the time is recalculated, that is, the time arrangement and seamless connection verification are re-executed on the adjusted scheme until the continuous executable condition is met.

[0134] VI. Based on the vehicle's spatiotemporal task sequence, closed-loop control is implemented for the vehicle, enabling the vehicle to arrive at the work node at the predetermined time and complete the task reception action, achieving seamless execution.

[0135] This solution is for vehicles. By task Move to the next task The driving process is controlled in a closed loop, and the task is set. The completion time is ,Task The start time is , Indicates the length of time the vehicle is used for transfer, within a time interval. Based on the vehicle's real-time path progress, position, speed, and acceleration, the system continuously calculates and applies control parameters to ensure the vehicle stays on track at all times. The conditions for receiving the call at the designated time have been met.

[0136] Assume the vehicle starts from the mission End node to task The planned driving path length at the starting node is Introduce progress coordinates along this path ,in Indicates that it is located in the task End node, Indicates arrival at the mission Starting node; vehicle At any moment The progress position is denoted as Speed ​​is denoted as Acceleration is denoted as To achieve "on-time arrival," a reference schedule for time calibration is constructed:

[0137] (15)

[0138] And define a reference speed:

[0139] (16)

[0140] Then, schedule error and speed error are defined:

[0141] (17)

[0142] The autonomous driving controller of the autonomous transport vehicle calculates and controls the acceleration in real time based on the error in each control cycle:

[0143] (18)

[0144] in , To control the gain, Given an acceleration limiting function, the vehicle motion satisfies:

[0145] (19)

[0146] When the arrival time If the following conditions are met:

[0147] (20)

[0148] Then determine the vehicle In the mission The starting node arrives on time and meets the acceptance conditions, triggering the task. The task is received and the next task control process begins; among which... To allow for tolerance at the destination location, For speed tolerance, This is the allowed target speed when accepting the task. If the acceptance conditions are not met, it is determined that there is a deviation in the current "task chain and timestamp" at the execution layer, and a rollback process is initiated: the vehicle's arrival point deviation and speed deviation are calculated, and the minimum time margin that needs to be compensated is calculated.

[0149] (twenty one)

[0150] Then the margin is fed back to the unified spatiotemporal constraint set to update subsequent constraints, and the time-level task arrangement is re-executed to correct the start / end time of subsequent tasks; if the risk of chain breakage cannot be eliminated after time re-arrangement, the connection relationship between adjacent tasks is further adjusted and the task chain is reconstructed; after obtaining the updated "task chain and timestamp", the vertical closed-loop control and judgment of this step are re-entered until the equation (20) is satisfied to achieve on-time pickup.

[0151] Therefore, by utilizing the "precise and controllable" execution characteristics of autonomous transport vehicles, the time-level arrangement results of "task chain and timestamp" are implemented as a timely arrival control mechanism for adjacent tasks at work nodes, so that seamless connection is not only limited to the planning layer output, but can be stably realized at the execution layer. Specifically, for the connection of adjacent tasks of the same vehicle at the node, this scheme takes "the next task must arrive at its starting node and have the conditions for acceptance" as the control objective, constructs a reference arrival process based on the available time and the travel distance between nodes, and performs closed-loop correction during vehicle travel: first, the "time arrival" requirement is mapped to the reference travel progress and reference speed through equations (15) and (16), the deviation of the vehicle's real-time progress / speed relative to the reference quantity is characterized by equation (17), and the acceleration adjustment quantity with limited amplitude constraint is generated by equation (18) to drive the vehicle state. The process of continuously approaching the reference process is finally determined by equation (20) at the beginning of the task to check the arrival error and the acceptance condition. If the determination is successful, the task acceptance is triggered and the next stage of connection control is entered. If the determination is unsuccessful, the deviation is quantified and fed back to the unified spatiotemporal constraint set, a new spatiotemporal task sequence is regenerated, and the above-mentioned on-time arrival control is reapplied until the trigger condition is met, forming a closed-loop link of "control execution - arrival determination - deviation feedback - spatiotemporal set reconstruction - re-arrangement - re-control", ensuring that the "seamless connection" arrangement scheme has continuous and feasible stability and robustness.

[0152] In summary, this solution addresses the key issue in the management and control of traditional manually driven transport vehicles—that "merely completing vehicle-cargo matching or task sequencing is insufficient to form a directly executable plan"—by proposing a spatiotemporal collaborative seamless orchestration and control method for autonomous transport vehicles. Based on task elements such as the start and end locations, time windows, and task duration, as well as operational elements such as vehicle availability and road network / channel travel time, a unified spatiotemporal constraint framework for a set of work points is established. Within this framework, not only are the matching relationships between vehicles and cargo / transport tasks generated, along with the task acceptance chain structure at each work point, but the specific start and completion times of transport tasks at each work point are also planned, forming a time-level transport plan with "task chains and timestamps." Finally, through closed-loop control, vehicles arrive at work nodes at predetermined times and complete task pickup actions, achieving seamless execution. Unlike existing technologies that typically focus on task allocation or sequencing, this solution significantly improves the feasibility of scheduling results, the stability of continuous task connection, and transportation efficiency. It also has good constraint compatibility and scenario applicability, and can be widely applied to various vehicle transportation organization scenarios such as port yard transshipment, short-distance transfer in industrial parks / mining areas, warehouse loading and unloading coordination, and urban delivery.

Claims

1. A seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization, characterized in that, Includes the following steps: S1. Obtain the vehicle set and task set, and combine them with the spatiotemporal constraints to construct the task operation point set and the spatiotemporal constraint set; S2. Under the set of spatiotemporal constraints, with the optimization objective of minimizing vehicle completion time and minimizing task gap time, the task sequence corresponding to each work point is generated by solving the problem. S3. Arrange the time sequence for the task sequence corresponding to each work point, calculate the start and finish times of each task in the task sequence, so that adjacent tasks at the work point are connected continuously on the time axis. S4. Perform seamless connection verification on adjacent tasks at the same work point to determine whether the task sequence at the work point is continuous and executable. If the verification fails, proceed to S5. If the verification succeeds, proceed to S6. S5. Adjust the local scheme for the task sequence corresponding to the job point that failed the verification, and then return to S3; S6. Based on the task sequence corresponding to all successfully verified work points, control the working status of each vehicle accordingly, and implement closed-loop control of the vehicle's driving process between work points. The vehicle set in S1 is specifically as follows: It includes several autonomous transport vehicles, each of which It has the following attributes: vehicle initial position Available start time and from any position Arrive at the location travel time ; The task set is specifically as follows It contains several tasks, any one of which It has the following attributes: starting point ,end Task duration and time window The time window represents the range of times during which a task can begin; The set of task operation points constructed in S1 is as follows: This includes multiple work points where vehicles pick up tasks; The spatiotemporal constraint set constructed in S1 is as follows: in, This indicates the task-vehicle matching constraint. This indicates a sequence constraint of work points. Indicates time window constraints. Represents spatial reachability constraints; To ensure that each task at each work point is assigned to one vehicle, and that the same vehicle performs only one task at any given time, it is specifically represented as follows: In the formula, Assign variables to the task, when the vehicle At the work site Execute the task hour ,otherwise ; This is used to ensure that each task is assigned to a corresponding job point, and that the task order at each job point must satisfy the timing constraint, specifically expressed as: , , , In the formula, where, Assign variables to tasks at work points if and only if the task Assigned to work site hour ; For task adjacency variables, if and only if at the job point Location, Task After being assigned, the task immediately followed. If also assigned, then ; As the first task execution variable, when the task... For work points When the first task is assigned, The relationship between the above variables is as follows: , ; This is used to ensure that the start time of each task must satisfy the time window constraint, specifically expressed as: In the formula, Indicates task The start time; This is used to ensure that vehicles must meet spatial accessibility constraints while performing tasks at work sites, and it is specifically expressed as follows: In the formula, Indicates task The completion time.

2. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 1, characterized in that, The specific optimization objective in S2 is as follows: in, This indicates the completion time of the last task. Indicates work point Task With the task The gap time between them; Specifically, S2 employs a genetic algorithm to solve the optimization objective under a set of spatiotemporal constraints, thereby generating a vehicle task matching sequence corresponding to each work point.

3. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 2, characterized in that, The process of solving the optimization objective using a genetic algorithm under the set of spatiotemporal constraints in S2 includes: First, an initial population is generated. Each chromosome in the population represents a vehicle-cargo matching scheme and task sequence arrangement. Each chromosome consists of two parts, including a representation of the job point. Task j Assigned to vehicles i Task assignment coding And indicates the same work point p Task sequence coding of adjacent connection relationships at the location ; The fitness of each chromosome in the initial population is then evaluated, and the fitness function is designed according to the optimization objective: Chromosomes with higher fitness indicate less vehicle completion and waiting time. Based on the fitness function value, selection, crossover, and mutation operations are performed to generate a new generation of the population. This new generation will continue to undergo fitness evaluation, selection, crossover, and mutation until the optimal solution that satisfies the spatiotemporal constraints and achieves the target fitness value is found. The final vehicle task matching sequence at each task operation point is represented as a set: For any work point : , It is a set, and the job points are given at the same time. p Relevant tasks j By vehicle i Acceptance, and adjacent tasks j and k The order in which they occur.

4. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 3, characterized in that, The specific process of S3 is as follows: First, calculate the start time of the first task for each work point: ,in, For work points The moment the first mission begins, For vehicles Available start time, It is a vehicle From its initial position To the work site Travel time required to reach the location; The subsequent calculation of the start time of the task is as follows: ,in, For work points Complete the task Next task At the beginning of Indicates task The time of receipt, Indicates that the vehicle is on the mission Location To the mission Location Required travel time; Through the above process, the start and end times of each task at each work point are obtained. Based on the task sequence obtained in S2, the corresponding time sequence is assigned to the tasks at each work point, resulting in a task chain that satisfies the consistency of time sequence, ensuring that the tasks at each work point are connected in time sequence.

5. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 1, characterized in that, The seamless connection verification of adjacent tasks at the same work point in S4 includes work point continuity verification, time window consistency verification, and vehicle arrival point reachability verification. Specifically, the work point continuity verification is for adjacent task pairs at the same work point, requiring that the work gap between the planned start time of the subsequent task and the completion time of the preceding task does not exceed a preset threshold, and avoiding overlapping work intervals, so as to ensure continuous operation at work points without resource conflicts. The time window consistency check specifically requires that for any task, the time window constraint in the spatiotemporal constraint set must be satisfied to ensure that the start time of the plan falls within the task's allowed time window. The vehicle reachability verification specifically targets each vehicle undertaking adjacent tasks. It must satisfy the work point sequence constraint and spatial reachability constraint to ensure that the vehicle can arrive at the work point and complete the task acceptance before the start time of its corresponding task, so that the work point continuity requirement can be actually executed.

6. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 1, characterized in that, The local scheme adjustment in S5 includes sequence adjustment, reallocation, and time fine-tuning processes.

7. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 6, characterized in that, The sequence adjustment specifically involves, under the premise of a set of spatiotemporal constraints, locally exchanging, inserting, or reconnecting the order of tasks at the same work point, so that the adjacent relationships that cause work gaps to exceed the threshold or work intervals to conflict are replaced with adjacent relationships that meet the requirements of continuity and non-conflict. The reassignment specifically refers to tasks that violate time windows or reachability constraints. Adjust its assignment variables And update the associated ones simultaneously. The variable allows the task to be transferred to a vehicle that can arrive at the corresponding work point at the scheduled time, in order to fill the work gap and restore continuous operation; Specifically, the time fine-tuning involves locally compressing or shifting the start / end times of related tasks at the same work point without violating the time window constraint, so that the work point continuity requirement and the timing consistency constraint are met simultaneously.

8. The seamless orchestration control method for vehicle transportation tasks based on spatiotemporal three-dimensional collaborative optimization according to claim 4, characterized in that, The specific process of S6 includes: For vehicles By task Move to the next task During the driving process, the task is set. The completion time is ,Task The start time is , Indicates the length of time the vehicle is used for transfer, within a time interval. Based on the vehicle's real-time path progress, position, speed, and acceleration, the system continuously calculates and applies control parameters to ensure the vehicle stays on track at all times. Meet the on-time access conditions and implement closed-loop control as follows: Assume the vehicle starts from the mission End node to task The planned driving path length at the starting node is Introduce progress coordinates along this path ,in Indicates that it is located in the task End node, Indicates arrival at the mission Starting node; vehicle At any moment The progress position is denoted as Speed ​​is denoted as Acceleration is denoted as ; To achieve "on-time arrival", a reference schedule with time calibration is constructed. Define reference speed Schedule error With speed error The autonomous driving controller of the autonomous transport vehicle calculates and controls the acceleration in real time based on the error in each control cycle: in , To control the gain, Given an acceleration limiting function, the vehicle motion satisfies: When the arrival time If the on-time pickup condition is met: Then determine the vehicle In the mission The starting node arrives on time and meets the acceptance conditions, triggering the task. The task is received and the next task control process begins; among which... To allow for tolerance at the destination location, For speed tolerance, The target speed allowed when accepting the task; If the conditions for acceptance are not met, it is determined that there is a deviation in the current "task chain and timestamp" at the execution layer, and a rollback process is entered. The arrival point deviation and speed deviation of the vehicle are calculated, the minimum time margin that needs to be compensated is calculated, and the minimum time margin is fed back to the spatiotemporal constraint set to update the subsequent constraints, and then the process returns to execute S2.

Images