A multi-agv task group dynamic reorganization and collaborative scheduling method and system

Abstract

The application relates to the technical field of warehouse logistics, and proposes a multi-AGV task group dynamic recombination and collaborative scheduling method and system. The method comprises the following steps: S1, constructing a global resource pool taking a single AGV as a minimum scheduling unit; S2, when a new task or a task abnormal event is monitored, temporarily establishing a collaborative task group and assigning a role; S3, generating a path time plan for each AGV in the group to meet the collision avoidance and synchronization constraints; S4, controlling the execution of the plan, and when an event that needs to be recombined is monitored, incrementally adjusting the task group composition and path in a manner of minimizing disturbance; and S5, dissolving the task group and updating the resource pool after the task is completed. The method forms a management closed loop through fine-grained resource management, event-triggered intelligent grouping, collaborative path planning and incremental adjustment, overcomes the problems of insufficient resource utilization and slow response caused by fixed grouping, and improves the operation efficiency, agility and robustness of the multi-AGV cluster in a dynamic environment.

Description

Technical Field

[0001] This application relates to the field of warehousing and logistics technology, and in particular to a method and system for dynamic reorganization and collaborative scheduling of multiple AGV task groups. Background Technology

[0002] In modern automated warehousing and logistics and flexible manufacturing systems, the collaborative operation of Automated Guided Vehicle (AGV) clusters is key to improving overall operational efficiency. To handle large or heavy materials, the industry commonly employs a working model where multiple AGVs are pre-assigned to fixed task groups or collaborative transport units. In this model, the AGVs within a group collaborate according to a pre-defined master-slave relationship to jointly complete the handling task. Existing technologies have developed numerous solutions for scheduling and managing these fixed task groups, such as methods for task allocation and path planning among multiple independent AGVs or pre-assigned groups using a central scheduler.

[0003] However, existing technologies treat the entire task group or vehicle as the smallest unit of resource management and scheduling. This coarse-grained management approach reveals significant flaws in real-world dynamic operating scenarios: when an AGV within a task group fails, the system cannot simply replace the faulty unit. Instead, it often needs to schedule another complete, idle task group to take over the task, leaving other normal AGV resources within the faulty group idle and potentially over-consuming the system's overall idle resources. On the other hand, when the system receives a complex new task exceeding the capacity of a single task group, it cannot intelligently and dynamically combine the dispersed AGV resources in different states (such as about to become idle or operating independently) in real time. The system can only passively wait or search for a complete, idle task group with matching capabilities, resulting in task response delays and rigid resource allocation.

[0004] Therefore, the existing technology is essentially a static resource management model based on pre-grouping, which lacks the ability to perceive and dynamically reorganize AGV resources. As a result, when the system faces equipment failure or complex tasks, resource allocation becomes rigid and response is slow, and the flexibility and efficiency of the overall operation are severely restricted. Summary of the Invention

[0005] In view of this, this application proposes a method and system for dynamic reorganization and collaborative scheduling of multiple AGV task groups, aiming to solve the technical problems of rigid resource allocation, slow response and insufficient system flexibility in the existing AGV scheduling method based on pre-grouping.

[0006] The technical solution of this application is implemented as follows: In a first aspect, this application provides a method for dynamic reorganization and collaborative scheduling of multiple AGV task groups, characterized by the following steps: S1. Real-time acquisition of status information of each automated guided vehicle in the system, constructing a global resource pool with a single automated guided vehicle as the smallest scheduling unit; and continuously monitoring system events. When an event that meets the preset dynamic reorganization triggering conditions is detected, the target task to be processed triggered by the event is determined. The status information includes at least location, task execution status and battery level, and the events include at least new task requests and existing task execution anomalies. S2. Based on the requirements and constraints of the target task to be processed, and combined with the real-time status information of each automated guided vehicle in the global resource pool, a temporary collaborative task group is dynamically formed and a collaborative role is assigned to each automated guided vehicle in the temporary collaborative task group. S3. Generate a path and time plan for each automated guided vehicle in the temporary collaborative task group, wherein the path and time plan satisfy collision avoidance and collaborative synchronization constraints. S4. Control the temporary collaborative task group to execute the path and time plan, and monitor the execution process; if an event that requires dynamic reorganization is detected, make incremental adjustments to the composition and path planning of the temporary collaborative task group based on the current execution status, and generate and execute the adjusted path and time plan.

[0007] In some embodiments, in step S2, the dynamic formation of a temporary collaborative task group based on the requirements and constraints of the target task and in conjunction with the real-time status information of each automated guided vehicle in the global resource pool includes the following steps: S21. Based on the type of the event, determine multiple decision factors including spatial distance, estimated arrival time, current load rate, remaining power, and task interruptibility. S22. Normalize the values ​​of each decision factor for each candidate automated guided vehicle. S23. Dynamically assign weights to the decision factors, wherein, in response to the event of an existing task execution anomaly, increase the weights of the spatial distance and estimated arrival time factors; in response to the event of a new task request, increase the weights of the current load rate and remaining power factors. S24. Calculate the overall matching degree of each candidate automated guided vehicle based on the weighted decision factors, and select a preset number of automated guided vehicles based on the matching degree to form the temporary collaborative task group.

[0008] In some embodiments, the existing task execution anomaly includes at least one of the following specific situations: AGV failure, task path continuous blockage for more than a preset duration, and actual task progress lagging behind planned progress for more than a preset threshold; the new task request includes a large material transportation task that requires multiple AGVs to work together, or a task optimization request where the system detects that the idle rate of AGVs in a local area exceeds a preset load balancing threshold.

[0009] In some embodiments, the step of assigning collaborative roles to each automated guided vehicle within the temporary collaborative task group includes: If the event is an abnormal execution of an existing task and involves replacing a faulty automated guided vehicle (AGV), the newly selected AGV will inherit the role of the replaced vehicle. If the event is a new task request, or an existing task execution exception that does not involve direct replacement, then based on the real-time status information of each automated guided vehicle, the one with the best status is designated as the master AGV, and the rest are designated as slave AGVs.

[0010] In some embodiments, step S3, which generates a path and time plan for each automated guided vehicle in the temporary collaborative task group, includes: S31. Using the A* algorithm, an initial path from the starting point to the target point is independently planned for each automated guided vehicle in the temporary collaborative task group; wherein, a weight coefficient based on task priority is introduced into the cost function of the A* algorithm to plan a lower-cost path for the automated guided vehicle executing the current target task. S32. Assign a time window to each path node on the initial path and perform conflict detection; S33. When a conflict is detected between the path time windows of different automated guided vehicles, based on the task priority associated with the path and time plan, a detour route is planned for the low-priority automated guided vehicles or a waiting time is inserted into their time windows to eliminate the conflict and form the final path and time plan.

[0011] In some embodiments, after S33, a path smoothing step is further included: without violating the determined time window and safety distance constraints, the final path after conflict elimination is smoothed using B-spline curves to reduce path turning points.

[0012] In some embodiments, the step of incrementally adjusting the composition and path planning of the temporary collaborative task group in a manner that minimizes system disturbance includes: If the path is blocked due to dynamic obstacles, or if a specific automated guided vehicle (AGV) in the temporary collaborative task group malfunctions, the composition and path planning of the temporary collaborative task group will be incrementally adjusted. The adjustment includes local path replanning for the affected AGVs and replacement and continuation path planning for the members of the temporary collaborative task group. When the adjustment is performed, the complete paths of the AGVs in the temporary collaborative task group that are not affected by the event will not be replanned.

[0013] In some embodiments, the local path replanning adopts a dynamic window method, which selects the trajectory with the optimal evaluation function from multiple sampled trajectories and issues it as a control command, taking into account the current speed, acceleration constraints, distance to dynamic obstacles, minimum safe distance and tracking deviation relative to the global path.

[0014] In some embodiments, when making the incremental adjustment, a historical adjustment strategy library that matches the current event type and the status of the affected automated guided vehicles is invoked to determine the adjustment scheme.

[0015] Secondly, this application discloses a dynamic reorganization and collaborative scheduling system for multiple AGV task groups, comprising: The global resource management module is used to acquire the status information of each automated guided vehicle (AGV) in the system in real time, construct a global resource pool with each AGV as the smallest scheduling unit, and continuously monitor system events. When an event that meets the preset dynamic reorganization trigger conditions is detected, the target task that needs to be processed is determined by the event. The status information includes at least location, task execution status, and battery level, and the events include at least new task requests and existing task execution anomalies. The intelligent decision-making module is used to dynamically form temporary collaborative task groups and assign collaborative roles to each automated guided vehicle in the temporary collaborative task group based on the requirements and constraints of the target task to be processed and combined with the real-time status information of each automated guided vehicle in the global resource pool. The collaborative planning module is used to generate path and time plans for each automated guided vehicle in the temporary collaborative task group that meet the constraints of collision avoidance and collaborative synchronization. The execution control module is used to control the temporary collaborative task group to execute the path and time plan, and to monitor the execution process. If an event that requires dynamic reorganization is detected, the composition and path planning of the temporary collaborative task group are incrementally adjusted based on the current execution state, and the adjusted path and time plan is generated and executed.

[0016] This application has the following advantages over the prior art: (1) The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups provided in this application achieves refined resource perception by establishing a global resource pool with a single AGV as the granularity; it achieves dynamic and flexible formation of temporary collaborative task groups through an event triggering mechanism and intelligent matching based on real-time status; it ensures the safety and orderliness of collaborative operations by applying collision avoidance and collaborative synchronization constraints in path planning; it achieves efficient and low-disturbance response to dynamic changes by monitoring and incremental adjustments during execution; and it forms a closed-loop management by disbanding the task groups and updating the resource pool. This method effectively overcomes the problems of low resource utilization, slow response, and poor flexibility caused by fixed grouping in the prior art, and significantly improves the overall operating efficiency, agility, and robustness of multi-AGV clusters in complex dynamic environments.

[0017] (2) By constructing a complete quantitative calculation process, the abstract decision of dynamically forming task groups is transformed into concrete and objective mathematical model calculations. This process clarifies multi-dimensional decision factors, achieves data comparability through normalization, dynamically allocates weights according to scenario objectives to reflect cost differences, and finally calculates the comprehensive matching degree through weighted summation and selects the best option accordingly. This mechanism enables the system to scientifically, automatically, and efficiently select the most suitable AGV combination to cope with the current emergency failure or new task requirements from the global resource pool, thereby realizing intelligent, refined, and scenario-based resource allocation at the technical level.

[0018] (3) The collaborative role assignment rules disclosed in this application provide a clear and efficient internal role allocation scheme for dynamically formed task groups through scenario-based and differentiated strategies. Role inheritance is adopted in fault replacement scenarios to maintain task continuity and reduce reorganization overhead; in new construction or structural adjustment scenarios, the optimal state leader is adopted to build the optimal intra-group collaborative architecture. This set of rules enables temporarily formed AGV teams to quickly form effective internal command and cooperation relationships, ensuring the order and efficiency of collaborative operations and supporting the agile and reliable operation of the entire dynamic reorganization scheduling method.

[0019] (4) The collaborative path planning steps disclosed in this application systematically solve the safety and efficiency problems in multi-AGV collaborative operations through a two-layer mechanism of priority-guided global path initial planning and time window coordination conflict resolution. The improved A algorithm injects priority advantages into key tasks at the planning source, while the time window-based conflict detection and resolution mechanism ensures the implementation of this priority at the execution level. By dynamically arranging detours or waiting for low-priority AGVs, a globally coordinated and locally conflict-free execution schedule is ultimately formed. This method enables temporarily formed AGV teams to achieve efficient, orderly, and reliable collaborative movement in a shared space, ensuring the smooth implementation of the entire dynamic reorganization scheduling scheme.

[0020] (5) By setting an incremental adjustment scheme that combines local path replanning with member replacement and continuation planning, and constraining it not to perform global replanning of the entire group's complete path, the system can quickly and accurately intervene only in the directly affected parts when encountering dynamic path blocking or vehicle anomalies. This scheme avoids global scheduling oscillations triggered by local problems, thereby significantly reducing the computational and communication overhead caused by the adjustment process, reducing the chain interference on the overall task execution process, and ensuring the high response speed, high execution continuity, and overall stability and efficiency of the multi-AGV system in dynamic environments. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of the dynamic reorganization and collaborative scheduling method for multiple AGV task groups disclosed in the embodiments of this application; Figure 2 This is a schematic diagram of the structure of the multi-AGV task group dynamic reorganization and collaborative scheduling system disclosed in the embodiments of this application; Figure 3 This is a schematic diagram of the device structure of the hardware operating environment of the electronic device disclosed in the embodiments of this application. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] like Figure 1 As shown, the first embodiment of this application discloses a method for dynamic reorganization and collaborative scheduling of multiple AGV task groups. The method mainly includes the following steps: S1. Real-time acquisition of status information of each automated guided vehicle in the system, constructing a global resource pool with a single automated guided vehicle as the smallest scheduling unit; and continuously monitoring system events. When an event that meets the preset dynamic reorganization triggering conditions is detected, the target task to be processed triggered by the event is determined. The status information includes at least location, task execution status and battery level, and the events include at least new task requests and existing task execution anomalies.

[0025] This step is fundamental to the entire method. Specifically, the system continuously collects key status information from each Automated Guided Vehicle (AGV). This status information includes at least location, task execution status, and battery level.

[0026] The acquired location information reflects the real-time position of the AGV in the working environment and is the basic basis for any scheduling and path planning. The task execution status (e.g., idle, executing, faulty, charging) directly indicates whether the AGV is currently available and what stage of operation it is in, while the power information is related to the AGV's endurance and the sustainability of task execution.

[0027] By continuously maintaining this fine-grained status information, the system no longer manages fuzzy task groups, but instead constructs a global resource pool with each individual AGV as the smallest manageable unit. This resource pool achieves panoramic and refined perception of all mobile resources within the system, providing precise data support for any subsequent flexible scheduling based on individual AGVs, fundamentally breaking through the traditional coarse-grained resource management model that uses fixed groups as indivisible units.

[0028] Simultaneously, the system continuously monitors system events in parallel. These events include at least new task requests and existing task execution anomalies. New task requests represent newly added external workloads, while existing task execution anomalies encompass unexpected situations such as AGV malfunctions within the group or task timeouts. Once an event that meets the preset dynamic reorganization trigger conditions is detected, the system immediately performs logical judgment to determine the target task that needs to be processed, triggered by the event. For new task requests, the target task is the new task itself; for existing task execution anomalies, the target task is the unfinished portion of the abnormal task. Through this step, the system clarifies the target task and provides a comprehensive global resource pool for subsequent intelligent decision-making.

[0029] Step S2: Based on the requirements and constraints of the target task to be processed, and combined with the real-time status information of each automated guided vehicle in the global resource pool, dynamically form a temporary collaborative task group and assign a collaborative role to each automated guided vehicle in the temporary collaborative task group.

[0030] This step is the decision-making stage of the entire method. The system focuses on the target task determined in the previous step, analyzing its specific requirements (such as the required number of AGVs, carrying capacity, destination, etc.) and constraints. Then, it comprehensively matches and intelligently calculates these requirements with the real-time status of each AGV in the global resource pool (such as location, power, whether it can be scheduled, etc.).

[0031] This process doesn't involve selecting from a pre-defined, fixed group. Instead, it dynamically filters out several AGVs best suited for completing the target task from all available individual AGV resources and dynamically assembles them into a temporary collaborative task group. Simultaneously, the system dynamically assigns collaborative roles to members within the group based on task requirements and the status of each AGV. For example, one AGV is designated as the master AGV leading coordination, while others act as follower AGVs. This creates a temporary work unit that is built on demand, has a clear structure, and well-defined responsibilities.

[0032] Step S3: Generate a path and time plan for each automated guided vehicle in the temporary collaborative task group, wherein the path and time plan satisfy collision avoidance and collaborative synchronization constraints.

[0033] After forming a temporary collaborative task group, a movement trajectory from the starting point to the destination needs to be planned for each AGV within the group. The generated path and time plan are not independent of each other; they must satisfy two key constraints: collision avoidance and collaborative synchronization constraints.

[0034] Collision avoidance constraints ensure that the path planned for each AGV does not conflict with the paths of other AGVs in space, while also avoiding static obstacles in the environment. This is a basic requirement for ensuring operational safety. Coordination and synchronization constraints require that the paths of multiple AGVs be coordinated in time. For example, multiple AGVs jointly transporting the same material must arrive at the docking point simultaneously, or they must stagger their entry into narrow passages rather than simultaneously. By generating path and time plans that satisfy both constraints, the system ensures that temporarily assembled AGV groups can move safely, orderly, and coordinatedly towards a common target point as a whole, laying the foundation for efficient collaborative operations.

[0035] S4. Control the temporary collaborative task group to execute the path and time plan, and monitor the execution process; if an event that requires dynamic reorganization is detected, make incremental adjustments to the composition and path planning of the temporary collaborative task group based on the current execution status, and generate and execute the adjusted path and time plan.

[0036] This step demonstrates the dynamic adaptability and robustness of the method. In this embodiment, after the temporary collaborative task group begins executing the predetermined plan, the system does not stop working but continuously monitors the execution status. This monitoring aims to promptly detect new emergencies, such as the appearance of unknown dynamic obstacles that block the original path, or a sudden malfunction of an AGV within the group. These could all become events that require triggering dynamic reorganization again.

[0037] Once such an event is detected, the system will not rigidly continue with the original, no-longer-feasible plan, nor will it simply and abruptly order the entire task group to stop and completely replan. Instead, it will incrementally adjust the composition and path planning of the temporary collaborative task group based on the current execution status (such as the current position of each AGV and the task progress). For example, it may locally modify the subsequent path of the affected AGV only to avoid obstacles, or simply replace the faulty AGV and plan a connecting path for it and its new partner. The core of this incremental adjustment lies in its localization and specificity, aiming to solve the problem with minimal change costs. After the adjustment is completed, the system quickly generates and executes the adjusted path and time plan, allowing the task to continue, thereby ensuring the system's efficient response and stable operation in the face of dynamic changes.

[0038] The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups provided in this application achieves fine-grained resource perception by establishing a global resource pool at the granularity of individual AGVs; it enables the dynamic and flexible formation of temporary collaborative task groups through an event-triggered mechanism and intelligent matching based on real-time status; it ensures the safety and orderliness of collaborative operations by applying collision avoidance and collaborative synchronization constraints in path planning; it achieves efficient and low-disturbance response to dynamic changes by monitoring and incremental adjustments during execution; and it forms a closed-loop management system through task group disbandment and resource pool updates. This method effectively overcomes the problems of low resource utilization, slow response, and poor flexibility caused by fixed grouping in existing technologies, and significantly improves the overall operating efficiency, agility, and robustness of multi-AGV clusters in complex dynamic environments.

[0039] In some embodiments, this application discloses a strategy process for dynamically forming temporary collaborative task groups, the steps of which are as follows.

[0040] S21. Based on the type of the event, determine multiple decision factors including spatial distance, estimated arrival time, current load rate, remaining power, and task interruptibility.

[0041] This step is the data preparation phase for decision-making calculations. The system selects five core dimensions to evaluate candidate AGVs based on event type.

[0042] Among them, spatial distance Through formula The calculation characterizes the current position of the candidate AGV. x i ,y i ) to the task trigger point ( x t ,y t The Euclidean geometric length of a given element is the fundamental factor affecting the response speed.

[0043] Estimated arrival time From the formula Estimate its current time t 0 Based on the maximum driving speed, it is superimposed The travel time provides a more accurate prediction of time costs.

[0044] Current load rate This is a value between 0 and 1, reflecting the saturation level of the AGV's current tasks. Remaining battery power. Similarly, a percentage value between 0 and 1 directly relates to the AGV's endurance potential for performing tasks. Task interruptibility. It is also an identifier between 0 and 1. A value of 1 indicates that the current task can be interrupted immediately in response to a higher priority scheduler, while a value of 0 indicates that it cannot be interrupted.

[0045] These five factors constitute a comprehensive index system for evaluating AGVs from four aspects: time and space cost, resource availability, energy sustainability, and scheduling flexibility.

[0046] S22. Normalize the values ​​of each decision factor for each candidate automated guided vehicle.

[0047] Since the dimensions and numerical ranges of the decision factors are quite different, normalization is required to eliminate the influence of dimensions and enable them to be compared and calculated on the same scale. The system uses the min-max normalization method to map each factor value to the [0,1] interval.

[0048] For spatial distance, the formula is used. The value is processed so that a larger value indicates a shorter distance. The estimated arrival time is determined using the formula... The value is processed so that a larger value indicates an earlier arrival. For the current load rate, the formula is used... The values ​​are processed so that a larger value indicates a lighter load. Since the remaining battery power and task interruptibility are already in the [0,1] range, their normalized values ​​are directly taken from their original values. , After this step, all factors are transformed into dimensionless, directional, and standardized parameters (the larger the value, the better).

[0049] S23. Dynamically assign weights to the decision factors, wherein, in response to the event of an existing task execution anomaly, increase the weights of the spatial distance and estimated arrival time factors; in response to the event of a new task request, increase the weights of the current load rate and remaining power factors.

[0050] This step demonstrates the intelligence and scenario adaptability of the decision-making strategy; the system assigns a weight vector to the five normalized decision factors. And satisfy The weight allocation is not fixed, but dynamically adjusted according to the core objective pursued by the event type. When the event is an anomaly in the execution of an existing task, the primary objective is to respond quickly to minimize task downtime. Therefore, the weight combination strategy will be biased towards spatial distance and expected arrival time, for example, by using... to give and Higher weighting. When the event is a new task request, the core objective is to ensure the long-term stability and sustainability of task execution. Therefore, the weighting strategy will tilt towards resource availability and endurance, for example, by using... to give and Higher weighting. This differentiated weighting strategy ensures that the decision-making model can accurately match the fundamental needs of the two different scenarios: "emergency response" and "optimization".

[0051] S24. Calculate the overall matching degree of each candidate automated guided vehicle based on the weighted decision factors, and select a preset number of automated guided vehicles based on the matching degree to form the temporary collaborative task group.

[0052] After completing the dynamic weight allocation, the system calculates the overall matching degree of each candidate AGV using a linear weighted model. The calculation formula is: This formula integrates all decision factors and their importance weights to calculate... The value ranges from [0,1]. A higher value indicates a higher degree of match between the overall state of the AGV and the target task requirements in the current specific scenario. Subsequently, the system ranks all candidate AGVs according to their overall matching degree. Sort the AGVs in descending order. Finally, based on the pre-defined number K of AGVs required for the target task, select K AGVs with the highest matching degree from the top of the sorted list. These selected AGVs constitute the optimal temporary collaborative task group dynamically assembled for the current target task.

[0053] By constructing a complete quantitative calculation process, the abstract decision of dynamically assembling task groups is transformed into concrete and objective mathematical model calculations. This process clarifies multi-dimensional decision factors, achieves data comparability through normalization, dynamically allocates weights based on scenario objectives to reflect cost differences, and finally calculates the comprehensive matching degree through weighted summation and selects the best option accordingly. This mechanism enables the system to scientifically, automatically, and efficiently select the most suitable AGV combination from the global resource pool to cope with current emergency failures or new task requirements, thereby achieving intelligent, refined, and scenario-based resource allocation at the technical level.

[0054] In some embodiments, existing task execution anomalies include at least one of the following specific situations: AGV malfunction, task path continuously blocked for more than a preset duration, and actual task progress lagging behind planned progress for more than a preset threshold.

[0055] This section defines three types of abnormal scenarios that require system emergency response and intervention. Among them, AGV failure refers to mechanical, electrical, or communication problems that prevent the automated guided vehicle from continuing to perform its intended task. This is the most direct and urgent abnormal situation and must be dealt with immediately to prevent the task from coming to a complete standstill.

[0056] If a task path is blocked for more than the preset time, it means that the planned path is blocked due to dynamic obstacles, other devices blocking the way, etc., and the duration of the blockage exceeds the system's tolerance threshold. This indicates that waiting may be ineffective and it is necessary to actively change the path or scheduling strategy.

[0057] When the actual progress of a task lags behind the planned progress by more than a preset threshold, it means that the execution efficiency of the task has not met expectations and there is an unacceptable deviation between the actual completion and the time schedule. This situation is caused by a variety of reasons, such as AGV performance degradation and frequent obstacle avoidance, and requires system intervention to optimize the execution process or allocate additional resources.

[0058] All three scenarios represent unexpected interruptions or delays in the established task flow. By clearly identifying these as signals that trigger dynamic reorganization, the system can automatically switch from the planned execution mode to the exception handling and recovery mode, thereby ensuring the robustness of the system and the reliability of task completion.

[0059] In some embodiments, a new task request may include a large material transport task that requires multiple AGVs to work together, or a task optimization request where the system detects that the idle rate of AGVs in a local area exceeds a preset load balancing threshold.

[0060] This section defines another major category of events that trigger dynamic reorganization from the perspectives of task origin and system optimization. Large material transportation tasks requiring collaborative handling by multiple AGVs are a typical example of new collaborative operation requirements. Their size, weight, or shape exceeds the carrying capacity of a single AGV, and must be completed by a temporary team of multiple AGVs. This directly triggers the need to build a new task group from scratch.

[0061] Task optimization requests detected by the system when the idle rate of AGVs in a local area exceeds the preset load balancing threshold demonstrate the system's proactive efficiency optimization capabilities. When monitoring detects that the proportion of idle AGVs in a certain area is too high, indicating uneven resource utilization and potential for optimization, the system can autonomously generate a task optimization request even without any new external tasks. The goal of this request is to reallocate these idle resources, such as dynamically teaming them to perform other auxiliary tasks like inventory checks and area inspections, or preparing to respond to potentially high-priority tasks, thereby achieving overall system load balancing and improved resource utilization. Recognizing this system-driven optimization intent as a legitimate new task request empowers the scheduling system to evolve from passive response to proactive optimization.

[0062] In some embodiments, this application discloses the steps of assigning collaborative roles to each automated guided vehicle within the temporary collaborative task group.

[0063] If the event is an abnormal execution of an existing task and involves replacing a faulty automated guided vehicle (AGV), the newly selected AGV will inherit the role of the replaced vehicle.

[0064] This rule addresses specific scenarios where partial replacement is required due to member failure during task execution. When an AGV in a temporary collaborative task group fails and needs to be replaced, to maximize task continuity and minimize coordination complexity caused by reorganization, the system arranges for a newly selected AGV joining the task group to directly inherit the original role of the failed AGV. If the failed AGV was originally the master AGV responsible for path planning and leading collaboration, the new AGV will take over as the new master AGV; if the failed AGV was originally a slave AGV performing follow-up collaboration, the new AGV will be integrated as a slave AGV.

[0065] The advantage of this role inheritance mechanism is that it avoids triggering a renegotiation and reassignment of role relationships across the entire group due to a change in a single member. Existing collaborative logic, communication links, and some planned paths within the task group remain stable. New members only need to integrate by following the inherited role interface, greatly simplifying the reorganization process, accelerating task recovery, and minimizing fault handling.

[0066] If the event is a new task request, or an existing task execution exception that does not involve direct replacement, then based on the real-time status information of each automated guided vehicle, the one with the best status is designated as the master AGV, and the rest are designated as slave AGVs.

[0067] This rule applies to scenarios involving the formation of entirely new task groups or structural adjustments to existing task groups that do not involve direct member replacements. In such cases, there are no inheritable historical roles, and the system needs to rebuild an efficient collaborative architecture for the task group. The rule requires dynamically designating the AGV with the best overall status within the group as the master AGV based on the real-time status information of each AGV. "Best overall status" here is a comprehensive evaluation, which can be determined based on indicators such as remaining battery power, performance score, communication quality, or current load in the status information described in step S1. The vehicle designated as the master AGV will be responsible for leading path tracking, coordinating synchronization within the group, and communicating with the central system. The remaining AGVs are assigned as slave AGVs, responsible for receiving instructions from the master AGV and executing precise following and collaborative actions.

[0068] This dynamic assignment mechanism based on real-time status ensures that in any newly formed or adjusted task group, the leadership role is always assumed by the healthiest, most capable, or most suitable AGV, thus providing the best internal organizational foundation for the efficient and stable operation of the task group.

[0069] The collaborative role assignment rules disclosed in this application provide a clear and efficient internal role allocation scheme for dynamically formed task groups through scenario-based and differentiated strategies. Role inheritance is used in fault replacement scenarios to maintain task continuity and reduce reorganization overhead; in new construction or structural adjustment scenarios, the optimal performer takes the lead to build the optimal intra-group collaborative architecture. These rules enable temporarily formed AGV teams to quickly establish effective internal command and collaboration relationships, ensuring the order and efficiency of collaborative operations and supporting the agile and reliable operation of the entire dynamic reorganization and scheduling method.

[0070] In some embodiments, this application discloses a specific technical solution for generating path and time plans for each automated guided vehicle in the temporary collaborative task group, the solution including the following steps.

[0071] S31, Adopt A * The algorithm independently plans an initial path from the starting point to the target point for each automated guided vehicle in the temporary collaborative task group; wherein, in A * The algorithm's cost function incorporates weight coefficients based on task priority to plan a lower-cost path for the automated guided vehicle (AGV) performing the current target task.

[0072] This step is the starting point for collaborative path planning, aiming to calculate a spatially feasible basic path for each AGV in the group. The system employs the classic A-search algorithm, which evaluates a function... f(n)=g(n)+h(n) To find the optimal path from the starting point to the target point, where g(n) From the starting point to the node n The actual cost, h(n) From node n The estimated cost to reach the destination.

[0073] In order to reflect the differences in importance of different tasks during the planning phase, this invention addresses standard A. * The algorithm's cost function has been improved by introducing weight coefficients based on task priority. ω(π) The improved cost function is: f(p)=ω(π)· (g(p)+h(p)) Weighting coefficients ω(π) The rule for assigning values ​​is: assign a smaller weight to the AGV that is currently performing the target task to be processed, for example... ω(1) =0.8, while assigning a larger weight to AGVs that are currently performing other non-triggered tasks in the system, for example... ω(2) =1.2). This improvement allows the algorithm to generate a lower path cost for high-priority AGVs when calculating path costs. f(p) Value. During the path search process, the algorithm tends to expand. f(p) The algorithm prioritizes nodes with lower values, ultimately planning a lower-cost path for high-priority AGVs based on the algorithm's evaluation criteria. This typically results in a shorter path with fewer turns, securing a passage advantage from the initial planning stage and laying the foundation for subsequent collaborative scheduling.

[0074] S32. Assign a time window to each path node on the initial path and perform conflict detection.

[0075] After generating initial spatial paths for each AGV, the system needs to expand and coordinate these paths over time to prevent dynamic conflicts. A time window describes the occupancy of path resources over time. The system calculates and allocates a time window for each path node p on each initial path. TW(p)= [ t start (p),t end (p) ].in, t start (p) The estimated arrival time of the AGV at this node is calculated based on the current time, the cumulative travel distance before reaching the node, and the AGV's maximum speed. The calculation formula is as follows: ; tend (p) This indicates the time when the AGV leaves the current node and heads to the next node, and its calculation formula is: ,in, For the current time, This represents the maximum speed of the AGV. From the starting point to the node distance, ) is a node Distance to the next node.

[0076] By assigning time windows to all nodes along all paths, the entire movement process of each AGV is precisely described in the spatiotemporal dimension as a series of continuous spatiotemporal blocks. Subsequently, the system performs conflict detection, checking whether the spatiotemporal paths of different AGVs intersect. Conflicts mainly manifest in two scenarios: firstly, node conflicts, where the path time windows of two or more AGVs predict they will occupy the same node on the map at the same time; secondly, edge conflicts or opposing conflicts, where the path time windows of two AGVs predict they will pass towards each other through the same narrow passage or edge within similar time periods. Through systematic conflict detection, all potential spatiotemporal competition points can be identified in advance.

[0077] S33. When a conflict is detected between the path time windows of different automated guided vehicles, based on the task priority associated with the path and time plan, a detour route is planned for the low-priority automated guided vehicles or a waiting time is inserted into their time windows to eliminate the conflict and form the final path and time plan.

[0078] This step is the core decision-making stage of collaborative scheduling, aiming to proactively resolve detected path conflicts. The resolution strategy strictly follows the right-of-way principle based on task priority. When a conflict occurs between AGVs of different priorities, the system will ensure that the original path of the higher-priority AGV (i.e., the AGV executing the current target task) remains unaffected.

[0079] For low-priority AGVs, the system employs two flexible adjustment strategies: First, spatial avoidance, which involves replanning a local path to bypass the conflict node or area. This is achieved by calling a path planning algorithm to calculate an alternative path for the low-priority AGV. Second, time avoidance, which involves inserting a waiting time Δt into the original path time window of the low-priority AGV. For example, before reaching the conflict node p, it is made to wait. Δt=t conflict -t start (p) The time, of which t conflictThe high-priority AGV occupies the end time of the conflict node, thus ensuring that the low-priority AGV enters the node only after the high-priority AGV has passed. By comprehensively using detour and waiting strategies, the system resolves all detected path conflicts one by one without affecting the execution of high-priority tasks. Finally, it generates a final path and time plan for each AGV in the temporary collaborative task group that is conflict-free in both time and space and has a clear execution order.

[0080] The collaborative path planning steps disclosed in this application systematically solve the safety and efficiency challenges in multi-AGV collaborative operations through a two-layer mechanism: priority-guided global initial path planning and time-window-based conflict resolution. The improved A algorithm injects priority advantages into critical tasks at the planning source, while the time-window-based conflict detection and resolution mechanism ensures the implementation of this priority at the execution level. By dynamically arranging detours or waiting for low-priority AGVs, a globally coordinated and locally conflict-free execution schedule is ultimately formed. This method enables temporarily assembled AGV teams to achieve efficient, orderly, and reliable collaborative movement within a shared space, ensuring the smooth implementation of the entire dynamic reconfiguration and scheduling scheme.

[0081] After steps S31 to S33, the system has generated a conflict-free final path and time plan for each AGV in the temporary collaborative task group. However, this path is composed of a series of discrete path nodes, and the connections between nodes are usually straight segments, which may result in many sharp turns or zigzag curves in the path. When the AGV travels along such a path, it needs to frequently decelerate, turn, and accelerate again at turning points, which not only increases mechanical wear and control complexity but also reduces overall operating efficiency and may cause the cargo it carries to sway.

[0082] To address this issue, the system introduces path smoothing. The core of smoothing is optimizing the path geometry without altering its established time-space constraints. Maintaining the defined time window and safety distance constraints is a fundamental principle that the smoothing process must adhere to. This means that the smoothed path must ensure that the AGV arrives at its originally designated critical spatial locations within the scheduled time. These locations are related to the nodes bound to the time window, and the smoothed path curve must still maintain a safe distance from obstacles and other AGV paths. Smoothing is not a complete replanning process, but rather an optimization of the travel trajectory while satisfying the original collaborative scheduling results (time window) and safety rules.

[0083] The system employs B-spline curves as the mathematical tool for smoothing. B-spline curves are parametric curves with local controllability and continuous smoothness. The system will use a series of path nodes contained in the final path after conflict elimination. pi =(x i ,y i ) These serve as control points for the B-spline curve. Through calculation, a smooth, continuous curve C(t) that passes through or approximates these control points can be generated. Compared to the original path connected by straight line segments, this B-spline curve eliminates sharp corners at the nodes and replaces sharp turns with a continuously varying curvature arc, making the entire path smooth and even.

[0084] Through this path smoothing step, the final execution path obtained by the AGV is not only coordinated and conflict-free in terms of spatiotemporal planning, but also smooth and efficient in terms of geometry. This allows the AGV to track the path at a smoother speed and with fewer acceleration and deceleration operations, thereby reducing mechanical stress and energy consumption, shortening the overall task execution time, and improving the accuracy and stability of motion control. Therefore, this step further taps into performance potential and optimizes the actual driving experience and operational efficiency of the AGV without affecting the results of previous intelligent scheduling and conflict coordination.

[0085] In some embodiments, this application discloses a method for incrementally adjusting the composition and path planning of the temporary collaborative task group in a manner that minimizes system disturbance. The method includes the following steps.

[0086] If the path is blocked due to dynamic obstacles, or if a specific automated guided vehicle in the temporary collaborative task group malfunctions, the composition and path planning of the temporary collaborative task group will be incrementally adjusted.

[0087] This section first defines two typical events that trigger incremental adjustments. The first type is path obstruction caused by dynamic obstacles. This refers to the occurrence of moving obstacles that were not anticipated in the initial planning during the execution of the task, such as temporary personnel entering the work area, other equipment, or scattered goods, which prevent the AGV from continuing to pass safely along the original path.

[0088] The second category involves anomalies occurring in specific AGVs within a temporary collaborative task group. These anomalies include, but are not limited to, sudden communication interruptions, reporting of mechanical fault codes, positioning errors, or a sharp drop in battery power, rendering them unable to continue performing their duties within the current task group. Both types of events share the characteristic of disrupting the task group's established execution plan, but do not necessarily require a complete overhaul of the original plan. Upon identifying these events during monitoring, the system immediately initiates an incremental adjustment process, aiming to restore task feasibility with minimal change costs.

[0089] The adjustment includes one or a combination of the following: partial path replanning for the affected automated guided vehicles; replacement and succession path planning for the members of the temporary collaborative task group, without replanning the complete path of all automated guided vehicles in the group.

[0090] This section defines two core operational modes for incremental adjustments, which the system can apply individually or in combination depending on the specific nature of the event. The first operation is to perform local path replanning for the affected AGVs, which is mainly applicable to scenarios where paths are blocked. When the path of one or more AGVs in a group is blocked by a dynamic obstacle, the system does not replan the global path for the entire task group, but only instructs the directly affected AGVs to recalculate a new local path near their current position that avoids the current obstacle and eventually returns to the original global path or target point. This type of local replanning can use the dynamic window method, which generates a safe and feasible obstacle avoidance trajectory in real time, taking into account the current speed, distance to the obstacle, and deviation from the global path.

[0091] The second operation involves replacing and planning follow-up paths for the members of the temporary collaborative task group. This is primarily applicable to scenarios where AGVs within the group malfunction. When an AGV fails, the system matches a new available AGV from the global resource pool and adds it to the task group to replace the failed member. Replacement is not simply a member exchange; the key lies in follow-up path planning. The system needs to plan a connecting path for the newly added AGV from its current position to the current execution position of the task group. It may also need to fine-tune the subsequent paths of other AGVs that have close collaborative coupling with the original failed AGV to ensure that the new AGV can smoothly and synchronously integrate into the ongoing task flow, rather than causing the entire task group to stop and wait.

[0092] The key constraint for minimizing system disturbance is that, during the incremental adjustment, the complete paths of all AGVs in the group are not replanned. Regardless of path obstruction or member anomalies, the system's response strategy is limited to the AGVs directly affected by the event and their related path segments. For other AGVs in the task group that are unaffected by the event and are still executing normally according to the original plan, the system will maintain their original paths and time schedules, and will not allow unnecessary intervention or replanning. This constraint prevents global scheduling oscillations triggered by local problems, greatly reducing the computational overhead, communication burden, and cascading interference to the overall system execution flow caused by the adjustment operation itself, ensuring that the system maintains high stability and execution efficiency when dealing with emergencies.

[0093] By implementing an incremental adjustment scheme that combines local path replanning with member replacement and continuation planning, and by constraining it to avoid global replanning of the entire group's complete path, the system can quickly and accurately intervene only in the directly affected parts when encountering dynamic path congestion or vehicle anomalies. This scheme avoids global scheduling oscillations triggered by local problems, thereby significantly reducing the computational and communication overhead of the adjustment process, minimizing cascading interference to the overall task execution flow, and ensuring high response speed, high execution continuity, and overall operational stability and efficiency of the multi-AGV system in dynamic environments.

[0094] In some embodiments, the local path replanning disclosed in this application uses a dynamic window method to achieve real-time and safe local obstacle avoidance and trajectory generation.

[0095] The implementation process of this method is as follows: First, based on the current speed and acceleration limitations of the AGV, the system determines a feasible speed sampling space. (v,w) ,in v For linear velocity, w Angular velocity. For each set of sampled velocities... (v,w) The algorithm simulates its trajectory within a short time window (e.g., 0.5 seconds) in the future.

[0096] Next, the system uses an evaluation function. G(v,w) To score each simulated trajectory, the evaluation function integrates several key factors and takes the following form: .in, H(v,w) The consistency between the direction of the trajectory endpoint and the direction of the local target point is evaluated; the smaller the angle, the higher the score. D(v,w) The evaluation track is based on the distance between the track and the nearest obstacle (including dynamic obstacles). The greater the distance and the safer the track, the higher the score. V(v, w) When evaluating the linear velocity of a trajectory, a speed that maintains high forward efficiency is preferred. GT(v,w) This is an evaluation sub-item that is particularly emphasized in this application. It measures the deviation between the simulated trajectory and the global reference path. The smaller the deviation, the higher the score. This ensures that the local obstacle avoidance behavior does not deviate significantly from the predetermined global task path. α,β,γ,δ It is a dynamic weighting coefficient, and its allocation prioritizes ensuring obstacle avoidance safety.

[0097] The system selects an evaluation function through calculation. G(v,w) The trajectory corresponding to the highest-scoring sampling speed is used as the result of local replanning and is immediately converted into control commands and sent to the AGV for execution.

[0098] Under the above technical solution, by explicitly adopting the dynamic window method and listing the specific constraints it considers, the precise algorithmic meaning of the local path replanning technique is defined, ensuring the real-time performance, safety, and compliance with the global task during incremental adjustment in obstacle avoidance.

[0099] In some embodiments, the incremental adjustment decision-making mechanism of this application has been intelligently enhanced by introducing a learning and matching strategy based on historical experience.

[0100] Specifically, when the system detects an event requiring incremental adjustment again, the decision engine does not rely solely on fixed rules. It first extracts the type characteristics of the current event and the real-time status of the affected AGVs, then matches these characteristics against records in the historical adjustment strategy library. The goal of the matching is to find the historical case most similar to the current situation. Upon successful matching, the system prioritizes the adjustment plan adopted in that historical case and verified as effective, or combines plans from multiple similar cases to generate or determine the specific plan for this incremental adjustment. For example, if the historical library shows that when similar communication failures occur at similar locations, replacing the AGV with the same model and planning a straight connecting path has a high success rate and short recovery time, then the system will prioritize this plan this time.

[0101] By introducing learning and matching strategies based on historical experience, incremental adjustment evolves from a response based on immediate computation into a decision-making process that incorporates the wisdom of historical experience. It improves the accuracy and reliability of adjustment plan formulation by utilizing successful patterns inherent in historical data, helps avoid repeating inefficient attempts, thereby accelerating adjustment decision-making and increasing the success rate and system efficiency of the overall dynamic restructuring strategy.

[0102] In some embodiments, the method disclosed in this application further includes step S5: after the task is completed, disband the temporary collaborative task group and update the global resource pool.

[0103] Specifically, once the target task is successfully completed, the temporary collaborative task group formed for that task is declared to have completed its mission and is subsequently disbanded, with each AGV within the group reverting to an independent resource entity. The system synchronously updates the status information of these AGVs in the global resource pool, such as updating their task status to idle and refreshing their latest location and battery level. This step ensures the real-time and accurate data in the resource pool, guaranteeing that these AGVs can be considered and utilized in subsequent dynamic reorganization decisions, thus forming a complete closed-loop management process from resource awareness, dynamic team formation, collaborative execution to resource release.

[0104] See attached document Figure 2 As shown, the second embodiment of this application discloses a dynamic reorganization and collaborative scheduling system for multiple AGV task groups, including: The global resource management module 10 is used to acquire the status information of each automated guided vehicle (AGV) in the system in real time, construct a global resource pool with each AGV as the smallest scheduling unit, and continuously monitor system events. When an event that meets the preset dynamic reorganization triggering conditions is detected, the target task that needs to be processed is determined by the event. The status information includes at least location, task execution status, and battery level, and the events include at least new task requests and existing task execution anomalies. The intelligent decision-making module 20 is used to dynamically form a temporary collaborative task group and assign collaborative roles to each automated guided vehicle in the temporary collaborative task group based on the requirements and constraints of the target task to be processed and combined with the real-time status information of each automated guided vehicle in the global resource pool. The collaborative planning module 30 is used to generate path and time plans for each automated guided vehicle in the temporary collaborative task group that meet the constraints of collision avoidance and collaborative synchronization. The execution control module 40 is used to control the temporary collaborative task group to execute the path and time plan, and to monitor the execution process. If an event that requires dynamic reorganization is detected, the composition and path planning of the temporary collaborative task group are incrementally adjusted based on the current execution state, and the adjusted path and time plan is generated and executed.

[0105] In some embodiments, an electronic device is provided in this application. (Refer to the appendix.) Figure 3 As shown, it includes a processor 101 and a memory 102; the memory stores computer programs, wherein the computer program 103 implements the above-mentioned dynamic reorganization and collaborative scheduling method of multi-AGV task groups when executed by the processor.

[0106] Specifically, processor 101 may include, for example, a general-purpose microprocessor, an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor may also include onboard memory for caching purposes. The processor may be a single processing unit or multiple processing units for performing different actions of the method flow according to embodiments of this application.

[0107] Memory 102 can be any medium capable of containing, storing, transmitting, propagating, or transmitting instructions. For example, memory can include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, instruments, or propagation media. Specific examples of memory include: magnetic storage devices such as magnetic tape or hard disk drives (HDDs); optical storage devices such as optical discs (CD-ROMs); and also random access memory (RAM) or flash memory; and / or wired / wireless communication links.

[0108] This application also provides a computer-readable medium storing a computer program that, when executed by a processor, implements the above-described method for dynamic reorganization and cooperative scheduling of multi-AGV task groups. This computer-readable medium may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into that device / apparatus / system. The aforementioned computer-readable medium carries one or more programs, which, when executed, implement the method according to the embodiments of this application.

[0109] According to embodiments of this application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wired, optical fiber, radio frequency signals, etc., or any suitable combination thereof.

[0110] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for dynamic reorganization and collaborative scheduling of multiple AGV task groups, characterized in that, Includes the following steps: S1. Real-time acquisition of status information of each automated guided vehicle in the system, constructing a global resource pool with a single automated guided vehicle as the smallest scheduling unit; and continuously monitoring system events. When an event that meets the preset dynamic reorganization triggering conditions is detected, the target task to be processed triggered by the event is determined. The status information includes at least location, task execution status and battery level, and the events include at least new task requests and existing task execution anomalies. S2. Based on the requirements and constraints of the target task to be processed, and combined with the real-time status information of each automated guided vehicle in the global resource pool, a temporary collaborative task group is dynamically formed and a collaborative role is assigned to each automated guided vehicle in the temporary collaborative task group. S3. Generate a path and time plan for each automated guided vehicle in the temporary collaborative task group, wherein the path and time plan satisfy collision avoidance and collaborative synchronization constraints. S4. Control the temporary collaborative task group to execute the path and time plan, and monitor the execution process; if an event that requires dynamic reorganization is detected, make incremental adjustments to the composition and path planning of the temporary collaborative task group based on the current execution status, and generate and execute the adjusted path and time plan.

2. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 1, characterized in that: In step S2, the dynamic formation of a temporary collaborative task group based on the requirements and constraints of the target task and in conjunction with the real-time status information of each automated guided vehicle in the global resource pool includes the following steps: S21. Based on the type of the event, determine multiple decision factors including spatial distance, estimated arrival time, current load rate, remaining power, and task interruptibility. S22. Normalize the values ​​of each decision factor for each candidate automated guided vehicle. S23. Dynamically assign weights to the decision factors, wherein, in response to the event of an existing task execution anomaly, increase the weights of the spatial distance and estimated arrival time factors; in response to the event of a new task request, increase the weights of the current load rate and remaining power factors. S24. Calculate the overall matching degree of each candidate automated guided vehicle based on the weighted decision factors, and select a preset number of automated guided vehicles based on the matching degree to form the temporary collaborative task group.

3. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 2, characterized in that: The existing task execution anomalies include at least one of the following specific situations: AGV failure, task path continuous blockage for more than a preset duration, and actual task progress lagging behind planned progress for more than a preset threshold; the new task requests include large material transportation tasks that require multiple AGVs to work together, or task optimization requests where the system detects that the idle rate of AGVs in a local area exceeds a preset load balancing threshold.

4. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 1, characterized in that: The step of assigning collaborative roles to each automated guided vehicle within the temporary collaborative task group includes: If the event is an abnormal execution of an existing task and involves replacing a faulty automated guided vehicle (AGV), the newly selected AGV will inherit the role of the replaced vehicle. If the event is a new task request, or an existing task execution exception that does not involve direct replacement, then based on the real-time status information of each automated guided vehicle, the one with the best status is designated as the master AGV, and the rest are designated as slave AGVs.

5. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 1, characterized in that: In step S3, the steps of generating path and time plans for each automated guided vehicle in the temporary collaborative task group include: S31, Adopt A * The algorithm independently plans an initial path from the starting point to the target point for each automated guided vehicle in the temporary collaborative task group; wherein, in A * The algorithm's cost function incorporates weight coefficients based on task priority to plan a lower-cost path for the automated guided vehicle (AGV) performing the current target task. S32. Assign a time window to each path node on the initial path and perform conflict detection; S33. When a conflict is detected between the path time windows of different automated guided vehicles, based on the task priority associated with the path and time plan, a detour route is planned for the low-priority automated guided vehicles or a waiting time is inserted into their time windows to eliminate the conflict and form the final path and time plan.

6. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 5, characterized in that: Following S33, a path smoothing step is also included: without violating the established time window and safety distance constraints, B-spline curves are used to smooth the final path after conflict elimination in order to reduce path inflection points.

7. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 1, characterized in that: The step of incrementally adjusting the composition and path planning of the temporary collaborative task group based on the current execution state includes: If the path is blocked due to dynamic obstacles, or if a specific automated guided vehicle (AGV) in the temporary collaborative task group malfunctions, the composition and path planning of the temporary collaborative task group will be incrementally adjusted. The adjustment includes local path replanning for the affected AGVs and replacement and continuation path planning for the members of the temporary collaborative task group. When the adjustment is performed, the complete paths of the AGVs in the temporary collaborative task group that are not affected by the event will not be replanned.

8. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 7, characterized in that: The local path replanning adopts the dynamic window method. Taking into account the current speed, acceleration constraints, distance to dynamic obstacles, minimum safe distance and tracking deviation relative to the global path, the trajectory with the optimal evaluation function is selected from multiple sampled trajectories and issued as a control command.

9. The method for dynamic reorganization and collaborative scheduling of multi-AGV task groups as described in claim 7, characterized in that: When making the incremental adjustment, the historical adjustment strategy library that matches the current event type and the status of the affected automated guided vehicles is invoked to determine the adjustment scheme.

10. A dynamic reorganization and collaborative scheduling system for multiple AGV task groups, characterized in that, include: The global resource management module is used to obtain the status information of each automated guided vehicle in the system in real time and build a global resource pool with a single automated guided vehicle as the smallest scheduling unit. The system continuously monitors events, and when an event that meets the preset dynamic reorganization triggering conditions is detected, the target task that needs to be processed is determined by the event. The status information includes at least location, task execution status and battery level, and the events include at least new task requests and existing task execution anomalies. The intelligent decision-making module is used to dynamically form temporary collaborative task groups and assign collaborative roles to each automated guided vehicle in the temporary collaborative task group based on the requirements and constraints of the target task to be processed and combined with the real-time status information of each automated guided vehicle in the global resource pool. The collaborative scale block is used to generate path and time plans for each automated guided vehicle in the temporary collaborative task group that satisfy collision avoidance and collaborative synchronization constraints. The execution control module is used to control the temporary collaborative task group to execute the path and time plan, and to monitor the execution process. If an event that requires dynamic reorganization is detected, the composition and path planning of the temporary collaborative task group are incrementally adjusted based on the current execution state, and the adjusted path and time plan are generated and executed.

Images