Stamping line tail AGV scheduling system

Abstract

The application provides a stamping line tail AGV scheduling system, relates to the field of intelligent manufacturing logistics scheduling, adopts a hierarchical centralized collaborative control architecture, sets a business layer module, a scheduling core layer module, an execution control layer module and a digital twin simulation layer module, internally arranges three customized strategies of main and auxiliary position dynamic priority distribution, flexible partition scheduling and double anti-locking traffic control in the scheduling core layer, realizes task differentiated priority distribution, operation area logical isolation control, space and time collaborative active anti-locking control, and completes strategy preposition verification and running state visual monitoring through digital twin simulation. The application can effectively adapt to the rigidity production rhythm of the stamping line tail, improve the multi-AGV collaborative scheduling efficiency and long-term operation stability, reduce the engineering landing debugging cost, and is suitable for the intelligent material carrying scene of the stamping workshop line tail.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing logistics scheduling, specifically to an AGV scheduling system for the tail end of a stamping line. Background Technology

[0002] In the automotive manufacturing industry, stamping is a core and leading process in vehicle production. The continuity and efficiency of material handling at the end of the stamping line directly determine the upper limit of the vehicle manufacturer's production capacity. Modern high-speed stamping lines typically maintain a high-frequency continuous operation of 12-18 SPM, employing a rigid production mode of alternating master and slave positions for packing. This places stringent requirements on the real-time performance of material handling, cycle time matching, and multi-vehicle coordination stability.

[0003] Traditional material handling at the tail end of stamping lines has long relied on manual forklift operation, which suffers from drawbacks such as high labor intensity, low efficiency, high safety risks, and poor cycle time matching, and can no longer meet the intelligent production needs of the modern automotive manufacturing industry. With the large-scale application of AGV technology, general-purpose AGV scheduling systems are gradually becoming a logistics upgrade solution, but they exhibit significant adaptability limitations in the special scenarios of narrow passages, high cycle times, and high coupling at the tail end of stamping lines. Insufficient cycle time matching capability: The existing industry-standard "First-Come, First-Served (FCFS)" and "Shortest Distance First" task allocation strategies only use order generation time or spatial distance as the sole allocation basis, without distinguishing the urgency of the main and auxiliary workstations. When multiple main and auxiliary workstations trigger full-box transfer tasks at the same time, it is very easy for high-urgency tasks at the main workstation to have no available AGVs to respond in time, while AGV resources are occupied by low-urgency tasks at nearby auxiliary workstations, causing unplanned downtime of the high-speed stamping line; Low efficiency of traffic flow control: Under the global free scheduling mode, AGVs can accept orders across workstations and regions without discrimination. There is no logical constraint on the operation boundary for the layout of the stamping line tail workstation, which can easily lead to problems such as disordered AGV traffic flow, serious imbalance in road network traffic flow distribution, and a significant increase in the intersection rate of driving paths, which further exacerbates the decline in system traffic efficiency. Poor scenario adaptability: Traditional path planning algorithms (such as standard A) The theoretical optimization of the multi-focus single algorithm (such as Dijkstra's algorithm) uses only spatial Euclidean distance as the sole path cost function, failing to incorporate dynamic variables such as real-time congestion waiting time and regional locking status. In multi-vehicle concurrent scenarios, this leads to all AGVs converging into the shortest one-way main road, causing irreversible head-on confrontations between AGVs traveling in and out of the warehouse within narrow channels, resulting in permanent deadlocks. When the core channel in front is occupied and locked, the AGVs behind cannot actively identify the congestion and detour, continuing to queue. This disconnect between theoretical research and engineering application threatens the continuous and stable operation of the production line. Summary of the Invention

[0004] The purpose of this invention is to provide a AGV scheduling system for the tail end of a stamping line, which realizes differentiated priority scheduling of master and slave positions, flexible zoned traffic control, and dual anti-deadlock collaborative control. At the same time, it is equipped with a digital twin simulation system to realize pre-verification, thereby improving the efficiency of system engineering implementation and operational stability.

[0005] To achieve the above objectives, the present invention employs the following technical solutions.

[0006] A stamping line tail AGV scheduling system adopts a hierarchical centralized collaborative control architecture, including a business layer module, a scheduling core layer module, an execution control layer module, and a digital twin simulation layer module. The business layer module is used to connect to the stamping production line warehouse management system, collect the filling status of the main and auxiliary workstation material boxes in real time, generate standardized handling orders according to the stamping production cycle, and send them to the scheduling core layer module. The scheduling core layer module is used to receive transportation orders and complete AGV task allocation, global path planning and multi-vehicle collaborative traffic control through built-in master and slave position dynamic priority allocation strategy, flexible partition scheduling strategy and dual anti-deadlock traffic management strategy. The execution control layer module is used to receive instructions from the scheduling core layer module, drive the AGV to complete the entire process of picking up, transferring, and releasing goods, and transmit the AGV's operating status and operation results back in real time. The digital twin simulation layer module is used to construct a digital twin scenario that matches the physical stamping line tail at a 1:1 scale, and to perform pre-simulation verification of the scheduling strategy and visualize monitoring of the operation status.

[0007] Furthermore, the specific execution logic of the dynamic priority allocation strategy for primary and secondary positions in the scheduling core layer module is as follows: Configure process priority weights and deadline parameters for each handling task. The primary location full-box task is assigned a high priority weight and a 300s hard deadline, while the secondary location full-box task is assigned a low priority weight and a 600s soft deadline. Construct a dynamic priority calculation function for the tasks. , In the formula, Let be the dynamic priority value of the i-th task. As the process priority weight for the task, The task has been waiting for time. Given the task deadline; construct the task-vehicle matching cost function: , In the formula, The total cost of assigning the i-th task to the j-th AGV. Let j be the distance traveled by the j-th AGV to the task pickup point. , The weighting coefficient is used; the system completes task-vehicle matching according to the principle of minimizing total cost, and high-priority primary tasks can be forcibly reclaimed from idle AGV resources for priority execution.

[0008] Furthermore, the flexible partitioning scheduling strategy of the scheduling core layer module is executed according to the following logic: The multiple main and auxiliary material frames at the end of the stamping line are logically divided into several independent work areas. Each work area corresponds to at least one set of main and auxiliary material frames. Dedicated operation AGVs are configured one-to-one with each work area, and at least one universal backup AGV for the entire area is also configured. The mapping relationship between work area, task, and AGV is established through feature attribute fields. Each task generated in the work area carries a corresponding group label, and the dedicated job AGV is only allowed to receive tasks in the corresponding group. During task allocation, the matching relationship of attributes is checked beforehand. The matching scheme is a valid solution only when there is an intersection between the feature attributes of the task and the AGV. The general-purpose standby AGV takes over the tasks in the corresponding work area only when the dedicated operation AGV has no idle capacity.

[0009] Furthermore, the dual deadlock prevention traffic management strategy of the scheduling core layer module is implemented with the following specific logic: Exclusive traffic control zones are designated at the throat passage and core intersection of the finished product warehouse, and a static area locking mechanism is adopted: when an AGV enters the boundary of the control zone, it applies for an exclusive area lock. During the locking period, other AGVs are prohibited from entering the control zone. The exclusive lock is released after the AGV has completely left the boundary of the control zone. A dynamic time-cost path planning mechanism is adopted, in which the path cost function is set as a weighted sum of spatial distance and expected waiting time: , In the formula, For path path The total cost, , These are the weighting coefficients. This is the path space distance normalization function. This is a normalization function for the expected path wait time; When the controlled area is locked and the expected waiting time exceeds the preset threshold, the system automatically triggers path replanning to guide the AGV to detour through the backup channel.

[0010] Furthermore, the business layer module integrates a warehouse scheduling simulation unit, which incorporates a phase difference staggered start algorithm and a sequential cyclic filling algorithm. The phase difference staggered start algorithm sets an independent delay start time for each picking area to achieve staggered triggering of multi-workstation tasks. The sequential cyclic filling algorithm simulates the relay operation logic of alternating packing at the main and auxiliary workstations of a single workstation, and automatically generates handling orders in a cyclical manner. The warehouse scheduling simulation unit also supports visualization of cargo box status, simulation parameter configuration management, and one-click switching between multiple operating conditions.

[0011] Furthermore, the execution control layer module interfaces with the AGV vehicle control system to perform hard blocking control on the AGV's picking and placing actions, and receives an automatic confirmation code after the action is completed to provide feedback on the action execution status; the execution control layer module also interfaces with the human-machine interaction terminal to realize manual task issuance, status query and abnormal intervention.

[0012] Furthermore, the system also includes a hardware support unit, which includes a scheduling server, an AGV on-board controller, a human-machine interface terminal, and a smart charging pile. The scheduling server adopts an industrial-grade server, configured with an I7-7700 processor, 16GB of memory, and a 256GB SSD, and can stably schedule up to 20 AGVs. The AGV on-board controller is adapted to forklift AGVs and supports the RCS robot control system. The human-machine interface terminal is an Android system PAD / PDA terminal, which supports on-site task management and status monitoring.

[0013] Furthermore, bidirectional communication between different levels is achieved via HTTPS and TCP / IP protocols, with instructions and status data transmitted in JSON format. Client identity verification and parameter validity verification mechanisms are configured, and communication latency is controlled within 150ms.

[0014] The advantages of this invention are: With strong process adaptability, a dynamic priority scheduling strategy is customized for the rigid process of alternating packing of main and auxiliary positions on the stamping line, which greatly shortens the average response time of the main position task and eliminates production line downtime caused by untimely handling from the root. The scheduling efficiency has been significantly improved. The flexible partitioned scheduling breaks down the global scheduling into local micro-circulations, effectively reducing the path conflict rate and the proportion of invalid empty runs of AGVs. The single vehicle operation cycle stably meets the production line cycle constraints. It has high operational stability, and the dual anti-deadlock mechanism achieves coordinated prevention and control in spatial and temporal dimensions. Multiple AGVs can run in parallel to achieve zero deadlock. Under long-term extreme working conditions, the task completion rate and system availability remain at a high level. The project is highly feasible, adopts a layered architecture and standardized communication interfaces, and is equipped with complete software and hardware configuration specifications. It can be quickly integrated with the existing stamping production line control system. At the same time, it can achieve pre-verification through digital twin simulation, which greatly reduces on-site debugging costs and downtime risks. Attached Figure Description

[0015] Figure 1This is a schematic diagram of the overall layered architecture of the scheduling system of the present invention; Figure 2 This is a schematic diagram of the information interaction architecture of the logistics system of the present invention; Figure 3 This is a schematic diagram of the traffic control area for the finished product warehouse of this invention; Figure 4 This is a schematic diagram of the communication architecture between the simulation system and the scheduling system of this invention. Detailed Implementation

[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0017] This embodiment is applied to the stamping workshop of an automobile OEM. The workshop has two parallel stamping production lines. Each production line has multiple sets of main and auxiliary material discharge frames at the end. The rated stamping frequency of the production line is 18 SPM, and the cycle time of a single material frame to fill the box corresponds to a rigid cycle time requirement of 300s. The workshop is equipped with 8 forklift-type AGVs, of which 6 are dedicated operation AGVs and 2 are backup AGVs for the entire area. A 5-meter-wide main logistics channel is set between the two production lines, and the finished product warehouse at the end has a funnel-shaped road network topology.

[0018] like Figure 1 As shown, the stamping line tail AGV scheduling system in this embodiment adopts a hierarchical centralized collaborative control architecture. From top to bottom, it consists of a business layer module, a scheduling core layer module, an execution control layer module, and a digital twin simulation layer module. Each layer interacts bidirectionally via a standardized industrial communication interface, forming a complete closed loop of task issuance, scheduling execution, and status feedback. The overall information interaction architecture of the logistics system is as follows: Figure 2 As shown, data flows from top to bottom through the warehouse management system in the business layer, the robot scheduling system in the scheduling layer, and the robot execution system in the execution layer, ultimately driving the AGV to complete the physical handling operation. The operation results are then transmitted back along the original link, realizing a closed loop of data throughout the entire process.

[0019] The business layer module interfaces with the existing LMS (Location Management System) in the workshop, collecting sensor signals from each material rack in real time via TCP / IP protocol to identify the three working states of the racks: empty, partially filled, and full. When any primary or secondary rack reaches full capacity, the business layer module automatically generates a standardized handling order. The order information includes the rack location number, picking coordinates, unloading coordinates, process priority attributes, and deadline parameters, and is uniformly distributed to the scheduling core layer module via HTTPS protocol. The business layer module also integrates a warehouse scheduling simulation unit. This unit incorporates a phase difference staggered start algorithm and a sequential cyclic filling algorithm, which can replace real production line signals to simulate continuous production conditions during system debugging and strategy verification. The phase difference staggered start algorithm sets an independent delay start time for each picking area, realizing staggered triggering of multi-workstation tasks and avoiding concentrated resource competition. The sequential cyclic filling algorithm simulates the relay operation logic of alternating packing at the main and auxiliary positions of a single workstation, which can automatically generate handling orders in a loop and support 72 hours of uninterrupted continuous simulation testing. This simulation unit also supports visual display of box status, local configuration of simulation parameters, and one-click switching of multiple working mode, providing a standardized test platform for iterative optimization of scheduling strategies.

[0020] The scheduling core layer module, as the core control hub of the system, is deployed on an industrial-grade scheduling server. It has built-in task allocation unit, path planning unit, and traffic control unit. Through the coordinated cooperation of the primary and secondary position dynamic priority allocation strategy, flexible partition scheduling strategy, and dual anti-deadlock traffic management strategy, it completes the entire process of AGV task allocation, global path planning, and multi-vehicle collaborative traffic control.

[0021] When implementing the dynamic priority allocation strategy for primary and secondary positions, the system first completes the basic parameter configuration, assigning process priority weights to all primary full-box tasks. Hard cutoff period For 300 seconds, configure the process priority weight (W_{pi}) of all sub-position full-box tasks to be 1, and set the soft deadline for sub-position tasks. The time is 600 seconds. After the task is generated, the system determines the time based on the task's waiting time. The priority value (P_i) of each task is updated in real time through a dynamic priority calculation function. The calculation formula is as follows: , As waiting time increases, task priority is increased accordingly to prevent low-priority secondary tasks from experiencing prolonged starvation. During the task allocation phase, the system iterates through all pending tasks and idle AGVs, calculating the total cost of each matching scheme using a task-vehicle matching cost function. The cost function is as follows: , in Take 0.8, The allocation is based on priority, with a threshold of 0.2. The system ultimately selects the matching scheme with the lowest total cost for execution. When there are unallocated high-priority primary tasks in the task pool and no idle AGVs, the system triggers a resource preemption mechanism, forcibly recalling the nearest AGV executing a low-priority secondary task, prioritizing the execution of the primary task. In this embodiment, the average response time of the primary task can be controlled within 2.1 seconds, and all primary tasks achieve zero timeout response during 24 hours of continuous operation.

[0022] When implementing the flexible zoning scheduling strategy, the system first completes the logical area division, logically dividing the multiple sets of main and auxiliary material frames at the end of the stamping line into several independent work areas. Each work area corresponds to at least one set of main and auxiliary material frames. Simultaneously, dedicated AGVs are configured one-to-one with the number of work areas, and at least one additional universal backup AGV is set up. In this embodiment, a total of 6 independent work areas are divided, matched with 6 dedicated AGVs, and 2 universal backup AGVs are set up. The system establishes a mapping and binding relationship between work areas, tasks, and AGVs by extending the "order type" feature attribute field in the data dictionary. Each handling order generated in each work area carries a corresponding group label, and the order acceptance attribute of the dedicated AGV matches the corresponding group label. During task allocation, the system performs a pre-verification of attribute matching. Only when the characteristic attributes of the task and the AGV intersect are the matching schemes deemed valid. The general-purpose standby AGV can only take over the transportation tasks in the corresponding work area when the dedicated operation AGV has no idle capacity. Thus, without changing the physical hardware layout, the orderly isolation of the work area is achieved through logical constraints, fundamentally preventing disorderly cross-regional boundary crossings and path intersection conflicts of AGVs.

[0023] Traffic control layout in the finished goods warehouse area, as follows Figure 3 As shown, mutually exclusive control zones are delineated around the throat passage and core stopping points. A dual anti-deadlock traffic management strategy achieves deadlock prevention through a combination of static locking and dynamic planning. The system first delineates several mutually exclusive traffic control zones at the throat passage of the finished goods warehouse and core road network intersections, implementing a static area locking mechanism: when an AGV reaches the boundary trigger point of a control zone, it automatically requests exclusive use of that zone from the traffic control unit. During the locking period, other AGVs are prohibited from entering the control zone. Once the AGV has completely left the control zone boundary, the system immediately releases the exclusive lock, preventing AGVs traveling in opposite directions from forming a deadlock in the narrow passage. Simultaneously, the system optimizes the cost function of path planning, adjusting the traditional single spatial distance cost to a weighted sum of spatial distance and expected waiting time. The cost function is as follows: , in Take 0.3, Choosing 0.7 highlights the priority of time cost in route selection. This is the path space distance normalization function. This is a normalized function for the expected path waiting time. When a controlled area is locked, causing the expected path waiting time to exceed a preset threshold, the system automatically triggers global path replanning, guiding the AGV to detour via an alternate lane. This dynamic traffic diversion proactively resolves congestion. In this embodiment, the congestion rate of the main lane in the finished goods warehouse decreased by more than 80%, and no traffic deadlock events occurred during 72 hours of continuous full-load operation.

[0024] The execution control layer module interfaces with the onboard control system of each AGV. After receiving navigation paths and action commands from the scheduling core layer, it drives the AGVs to sequentially complete the entire process of picking up, transferring, placing, and returning to standby. For picking up and placing actions, the execution control layer implements hard-block control on the AGVs to ensure that the vehicles are absolutely stationary during the action execution. After the action is completed, the execution control layer receives an automatic confirmation code from the AGV, with confirmation code 1010 for picking up and confirmation code 1020 for placing, thus providing feedback on the action execution status. The execution control layer also collects the AGV's position, speed, remaining battery power, operating status, and alarm information in real time and transmits it back to the scheduling core layer and the business layer. The status data update frequency meets the real-time requirements of industry. At the same time, the execution control layer interfaces with the mobile human-machine interaction terminal in the workshop to realize manual task issuance, order execution progress query, and abnormal alarm intervention, adapting to the needs of on-site manual intervention.

[0025] The digital twin simulation layer module constructs a high-fidelity digital twin scene that matches the physical stamping line tail at a 1:1 scale. During the scene modeling stage, the on-site CAD base map is imported to complete the road network topology construction, functional area configuration, and coordinate calibration, ensuring that the virtual scene coordinates are aligned with the physical site at the millimeter level. At the same time, digital twins are constructed for each AGV, and physical dimensions, kinematic parameters, and safety envelopes consistent with the real vehicle are configured. The performance deviation between the simulation and the real vehicle is controlled within a reasonable range through kinematic calibration.

[0026] The communication architecture between the simulation system and the scheduling system is as follows: Figure 4 As shown, a three-layer communication architecture is adopted, consisting of a simulation system, a backend server, and an RDS scheduling system. Bidirectional data interaction is achieved via HTTPS, while all instructions and status data are transmitted in JSON format. Client identity verification and parameter validity checks are configured, and communication latency is controlled within the hundreds of milliseconds. The digital twin simulation layer supports multi-speed simulation operation, enabling production line cycle time matching tests, multi-vehicle collaborative scheduling tests, multi-factor orthogonal experiments, and long-term stability tests. All optimizations and adjustments to scheduling strategies are verified in the simulation environment before being implemented in the physical system. This allows for strategy iteration and performance verification without interrupting on-site production, while also enabling visualized monitoring of AGV operating status.

[0027] After the scheduling system in this embodiment is put into operation, the average cycle of a single AGV round trip is much shorter than the production line cycle time constraint. There is no deadlock or collision during the parallel operation of multiple AGVs, and the task completion rate reaches 100%. Under long-term full-load operation, the system availability rate remains stable at a high level, and all performance indicators meet the rigid requirements of continuous production in the stamping workshop. It can completely replace the traditional manual forklift transfer mode and effectively improve the automation level and operational stability of the logistics at the end of the stamping line.

[0028] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A stamping line tail AGV scheduling system, characterized in that, A hierarchical centralized collaborative control architecture is adopted, including a business layer module, a scheduling core layer module, an execution control layer module, and a digital twin simulation layer module; The business layer module is used to connect to the stamping production line warehouse management system, collect the filling status of the main and auxiliary workstation material boxes in real time, generate standardized handling orders according to the stamping production cycle, and send them to the scheduling core layer module. The scheduling core layer module is used to receive transportation orders and complete AGV task allocation, global path planning and multi-vehicle collaborative traffic control through built-in master and slave position dynamic priority allocation strategy, flexible partition scheduling strategy and dual anti-deadlock traffic management strategy. The execution control layer module is used to receive instructions from the scheduling core layer module, drive the AGV to complete the entire process of picking up, transferring, and releasing goods, and transmit the AGV's operating status and operation results back in real time. The digital twin simulation layer module is used to construct a digital twin scenario that matches the physical stamping line tail at a 1:1 scale, and to perform pre-simulation verification of the scheduling strategy and visualize monitoring of the operation status.

2. The stamping line tail AGV dispatching system according to claim 1, characterized in that, The specific execution logic of the primary and secondary dynamic priority allocation strategy of the scheduling core layer module is as follows: Configure process priority weights and deadline parameters for each handling task. The primary location full-box task is assigned a high priority weight and a 300s hard deadline, while the secondary location full-box task is assigned a low priority weight and a 600s soft deadline. Construct a dynamic priority calculation function for the tasks. ， In the formula, Let be the dynamic priority value of the i-th task. As the process priority weight for the task, The task has been waiting for time. The deadline for the task; Construction task - vehicle matching cost function: ， In the formula, The total cost of assigning the i-th task to the j-th AGV. Let j be the distance traveled by the j-th AGV to the task pickup point. , The weighting coefficient is used; the system completes task-vehicle matching according to the principle of minimizing total cost, and high-priority primary tasks can be forcibly reclaimed from idle AGV resources for priority execution.

3. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, The flexible partitioning scheduling strategy of the core scheduling layer module is specifically executed as follows: The multiple main and auxiliary material frames at the end of the stamping line are logically divided into several independent work areas. Each work area corresponds to at least one set of main and auxiliary material frames. Dedicated operation AGVs are configured one-to-one with each work area, and at least one universal backup AGV for the entire area is also configured. The mapping relationship between work area, task, and AGV is established through feature attribute fields. Each task generated in the work area carries a corresponding group label, and the dedicated job AGV is only allowed to receive tasks in the corresponding group. During task allocation, the matching relationship of attributes is checked beforehand. The matching scheme is a valid solution only when there is an intersection between the feature attributes of the task and the AGV. The general-purpose standby AGV takes over the tasks in the corresponding work area only when the dedicated operation AGV has no idle capacity.

4. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, The dual deadlock prevention and traffic management strategy of the scheduling core layer module is specifically executed as follows: Exclusive traffic control zones are designated at the throat passage and core intersection of the finished product warehouse, and a static area locking mechanism is adopted: when an AGV enters the boundary of the control zone, it applies for an exclusive area lock. During the locking period, other AGVs are prohibited from entering the control zone. The exclusive lock is released after the AGV has completely left the boundary of the control zone. A dynamic time-cost path planning mechanism is adopted, in which the path cost function is set as a weighted sum of spatial distance and expected waiting time: ， In the formula, For path path The total cost, , These are the weighting coefficients. This is the path space distance normalization function. This is a normalization function for the expected path wait time; When the controlled area is locked and the expected waiting time exceeds the preset threshold, the system automatically triggers path replanning to guide the AGV to detour through the backup channel.

5. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, The business layer module integrates a warehouse scheduling simulation unit, which incorporates a phase difference staggered start algorithm and a sequential cyclic filling algorithm. The phase difference staggered start algorithm sets an independent delay start time for each picking area to achieve staggered triggering of multi-workstation tasks. The sequential cyclic filling algorithm simulates the relay operation logic of alternating packing at the main and auxiliary workstations of a single workstation, and automatically generates handling orders in a cyclical manner. The warehouse scheduling simulation unit also supports visualization of cargo box status, simulation parameter configuration management, and one-click switching between multiple operating conditions.

6. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, The execution control layer module interfaces with the AGV vehicle control system to perform hard blocking control on the AGV's picking and placing actions, and receives an automatic confirmation code after the action is completed to provide feedback on the action execution status. The execution control layer module also interfaces with the human-machine interaction terminal to realize manual task issuance, status query and abnormal intervention.

7. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, The system also includes a hardware support unit, which comprises a scheduling server, an AGV onboard controller, a human-machine interface terminal, and a smart charging pile. The scheduling server is an industrial-grade server configured with an I7-7700 processor, 16GB of memory, and a 256GB SSD, capable of stably scheduling up to 20 AGVs. The AGV onboard controller is compatible with forklift AGVs and supports the RCS robot control system. The human-machine interface terminal is an Android system PAD / PDA terminal that supports on-site task management and status monitoring.

8. The AGV scheduling system for the tail of a stamping line according to claim 1, characterized in that, Each level communicates bidirectionally via HTTPS and TCP / IP protocols. Commands and status data are transmitted in JSON format. Client identity verification and parameter validity verification mechanisms are configured, and communication latency is controlled within 150ms.

Images