Heavy-load AGV transportation system based on central scheduling and scheduling method thereof

Abstract

The invention relates to a heavy object transportation system, in particular to a heavy-load AGV transportation system based on central scheduling and a scheduling method thereof, and aims to solve the problem that an existing method easily causes increase of fault risk, falling into a local optimal trap and sharp increase of cost, so that actual demand is difficult. The system comprises a central scheduling system which is in communication connection with an upper control system, and a plurality of heavy-load AGVs which are uniformly scheduled by the central scheduling system, the central scheduling system comprises a storage module, a task processing module, a control module, a communication module and a monitoring module. The method comprises the steps of building a system and setting a working area, the method comprises the steps of establishing a global path knowledge base, putting a heavy load AGV into a working area, splitting a job task into ordered atomic action sequences, selecting out a heavy load AGV group capable of executing the atomic action sequences, generating a control instruction, issuing the control instruction and returning job task execution progress state information. According to the invention, through central unified scheduling, the multi-vehicle linkage effect is improved.

Description

Technical Field

[0001] This invention relates to a heavy-duty transport system, specifically a heavy-duty AGV transport system based on central dispatch and its dispatching method. Background Technology

[0002] In the fields of automated logistics and intelligent manufacturing, Automated Guided Vehicles (AGVs) serve as key execution equipment. Their core function is to achieve unmanned material handling through pre-set paths or navigation systems, thereby improving operational flexibility and efficiency. With the deepening of industrial applications, a variant of AGV with significantly greater load-bearing capacity—the heavy-duty AGV—is increasingly becoming a rigid demand in industries such as automotive manufacturing, aerospace, ports, and heavy equipment. The industry typically defines AGVs with a rated load of at least 2 tons as heavy-duty AGVs, with load capacities reaching tens or even hundreds of tons, used to solve the problem of precise transfer of large and overweight materials. Compared to conventional AGVs, heavy-duty AGVs not only exhibit a significant increase in physical size and load capacity, but also bring a series of qualitatively different technical challenges, including extremely high control precision requirements (often reaching millimeter-level), enormous motion inertia, and complex multi-vehicle collaborative operations.

[0003] Currently, traditional scheduling systems designed for light-duty AGVs typically employ a distributed, weakly centralized control architecture. However, this architecture has revealed fundamental limitations and inadequacies when dealing with the demanding operating conditions unique to heavy-duty AGVs, primarily manifested in the following three core contradictions: First, there is the contradiction between the catastrophic consequences of runaway risks and the insufficient safety of distributed control. Heavy-duty AGVs transport valuable materials, and their enormous kinetic energy means that even a slight collision, deviation, or loss of synchronization can lead to major accidents such as equipment damage and production line paralysis. Especially when multiple heavy-duty AGVs are collaboratively transporting the same ultra-large component (such as an aircraft fuselage or wind turbine blade), absolute synchronization in speed, direction, and phase is required, forming a "virtual rigid platform." Traditional distributed scheduling relies on individual vehicle autonomous decision-making and local sensor information, making it difficult to achieve precise coordination and conflict prediction at the global level, and thus unable to provide deterministic safety guarantees for such high-risk operations.

[0004] Secondly, there is the contradiction between optimal system-level operational efficiency and the blindness of individual vehicle's local decision-making. In complex work area path networks, the simultaneous operation of multiple heavy-duty AGVs can easily cause congestion or even system deadlock at key geographical bottlenecks such as intersections and narrow passages. In the distributed scheduling mode, each AGV makes a decision based on its own optimal path (such as the shortest path), and uses the A* algorithm (an efficient heuristic search algorithm) or its variants for path planning. This often leads the entire system into the "tragedy of the commons," causing a significant decrease in overall transportation throughput.

[0005] Finally, there is the contradiction between the high complexity of single-vehicle control and the need for cost reduction and efficiency improvement at scale. To achieve high-load and high-precision movement, heavy-duty AGVs often adopt complex configurations with multiple wheel systems and multiple drive motors, and their kinematic models, power distribution, and precision control algorithms are extremely complex. If such complex algorithms and programs are fully loaded into the on-board controller, it will lead to a surge in the hardware cost (high-performance computing unit, high-end sensors) and software development difficulty of a single AGV.

[0006] In conclusion, given the unique characteristics of heavy-duty AGVs, developing a system and method capable of achieving refined, centralized, and unified scheduling has become crucial for overcoming industry bottlenecks and unleashing their full application potential. Summary of the Invention

[0007] The purpose of this invention is to address the problem that in existing distributed AGV scheduling methods, the autonomous decision-making of each heavy-duty AGV easily leads to increased failure risk, getting trapped in local optima, and soaring costs, making it difficult to meet the high-precision and safe transportation requirements of large quantities of heavy goods. The invention provides a heavy-duty AGV transportation system and scheduling method based on central scheduling.

[0008] To achieve the above objectives, the technical solution provided by this invention is: A heavy-duty AGV transportation system based on central dispatch is unique in that: This includes a central dispatching system that communicates with the upper control system, and several heavy-duty AGVs that are uniformly dispatched by the central dispatching system; The central dispatch system includes a storage module, a task processing module, a control module, a communication module, and a monitoring module. The storage module is used to store a device action information database and a global path knowledge base. The action information database contains control parameters corresponding to all controllable atomic actions of the heavy-load AGV; the global path knowledge base contains path information pre-calculated based on navigation points on the work area map. The task processing module is communicatively connected to the upper control system and is used to receive and parse the work tasks from the upper control system, break down the work tasks into an ordered sequence of atomic actions, and return the work task execution progress status information to the upper control system. The control module is connected to the storage module and the task processing module respectively, and is used to issue control commands to the specified heavy-duty AGV based on the device action information database and global path knowledge database in the storage module, and the ordered atomic action sequence obtained by the task processing module. The communication module is connected to the control module and to each heavy-duty AGV, and is used to send the control commands to the specified heavy-duty AGV and receive the status signals returned by each heavy-duty AGV. The monitoring module is connected to the communication module and is used to monitor the equipment body status information, real-time position and motion status information, and task execution progress status information of the heavy-duty AGV according to the status signal.

[0009] Furthermore, the equipment action information database is stored in a horizontally partitioned database format, which includes a point table, a line table, an equipment information table, and a neighboring point relationship table. The point table is used to store the static attributes of each navigation point in the map, including point number, X-coordinate, Y-coordinate, orientation angle, and point status; The line table is used to store the attributes of the path segment connecting two navigation points, including the path segment number, start number, end number, path length, preset empty speed, preset loaded speed, and path direction attribute. The equipment information table stores the static attributes and dynamic status of each heavy-duty AGV; the static attributes include equipment number, equipment name, MAC address, load capacity, size information, speed and acceleration; The nearest point relationship table is used to store pairs of navigation points with direct connectivity in the working area map to represent the topology of the path network.

[0010] Furthermore, the control instructions include movement control instructions and equipment operation control instructions; for atomic actions involving movement, the control module selects a path for the specified heavy-duty AGV from the global path knowledge base, and generates movement control instructions by combining the corresponding control parameters queried from the equipment action information base; for atomic actions not involving movement, the control module queries the corresponding control parameters from the equipment action information base to generate equipment operation control instructions.

[0011] Furthermore, the data communication protocol between the communication module and each heavy-duty AGV is the Modbus protocol, and the polling period is less than or equal to 100ms.

[0012] Furthermore, the device status information includes battery level, battery health, voltage and current during charging, and device fault codes; The real-time position and motion status information includes X-axis coordinates, Y-axis coordinates, real-time speed, heading angle, and motion control mode; The task execution progress status information includes the currently executing task number, the atomic action instructions being executed under the task, and the completion status of the atomic actions.

[0013] Meanwhile, the present invention also provides a scheduling method for the above-mentioned heavy-duty AGV transportation system based on central scheduling, which is characterized by including the following steps: Step 1: Build a heavy-duty AGV transportation system based on central dispatch, set up work areas and establish a global path knowledge base for these areas; Step 2: Deploy several heavy-duty AGVs into the work area; Step 3: Start working. The upper control system sends the task to the central dispatch system. The task processing module of the central dispatch system breaks down the task into an ordered sequence of atomic actions. Step 4: The control module traverses all heavy-load AGVs through the communication module and selects and designates one or more heavy-load AGVs that can execute the above atomic action sequence; Step 5: For each atomic action in the atomic action sequence, query the corresponding control parameters from the device action information database; if the atomic action involves movement, perform path planning for each specified heavy-duty AGV from the global path knowledge base established in Step 1; generate control instructions based on the queried control parameters and the planned path. Step 6: Send the control command to each designated heavy-duty AGV to drive it to execute the atomic action sequence in sequence. During and after execution, the task processing module returns the task execution progress status information to the upper control system.

[0014] Furthermore, step 1 specifically includes: Step 1.1: Define the working area. Based on the line table and nearest-neighbor relationship table in the equipment motion information database, establish a directed weighted graph G={V,E,W}, where V is the vertex set, E is the edge set, and W is the weight matrix; V={v1,v2…v n}, where n is the number of navigation points in the working area, v1, v2…v n Representing different navigation points; for any edge e in the edge set E, represented by the triple (v i ,v j ,w ij ) indicates that v i and v jLet w represent the starting and ending navigation points of edge e, respectively. ij This indicates that the heavy-load AGV starts from the initial navigation point v. i Drive directly to the final navigation point v j The cost of passage; the w ij The elements that constitute the corresponding positions in the weight matrix W; if there is no direct path between two navigation points, then w is defined. ij =∞, and the travel cost w from a navigation point to itself ii =0; Step 1.2: Initialize an n×n travel cost matrix dist. Let the initial travel cost matrix dist... (0) =Weight matrix W, defined dist (0) [i][j]=w ij The initial passage cost matrix dist (0) As the initial state for iteration; Step 1.3: Sequentially select each navigation point in the vertex set V as a candidate transit navigation point v. k For the initial passage cost matrix dist (0) Begin the iteration, where k = 1, 2, ..., n; During each iteration, for each navigation point pair (v) i ,v j Perform the following operations: a. Calculation toll costs ,in This is the toll cost matrix after the previous iteration; b. Based on and The corresponding path determines the starting navigation point v. i transit navigation point v k To the final navigation point v j Complete vertex sequence ; c. Path information tuple Add to navigation point pair (v i ,v j The corresponding set of alternative paths middle; Step 1.4: After completing all iterations, based on each navigation point pair (v) i ,v j ), and its set of alternative paths All paths are sorted in ascending order based on their travel cost; the top five paths with the lowest travel cost are selected and stored in the storage module as a global path knowledge base.

[0015] Furthermore, in steps 1.1-1.4, the travel cost is distance.

[0016] Furthermore, in step 5, path planning is performed for each specified heavy-duty AGV from the global path knowledge base, specifically as follows: a1. Based on the starting point and ending point of the current atomic action, obtain from the global path knowledge base five pre-stored alternative paths and their distances for the navigation point pair formed by the starting point and ending point. a2. Check the real-time occupancy status of the five candidate paths, and select the path with the shortest distance among the unoccupied paths as the execution path.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. The heavy-duty AGV transportation system based on central scheduling provided by this invention establishes a directed weighted graph G based on the work area after setting the work area. Then, it adopts a variant of the Floyd-Warshall algorithm. This variant algorithm records and maintains a candidate set containing multiple better paths during the iteration process, thereby pre-calculating and storing a multi-path decision set (such as the top five optimal paths) sorted by cost for any pair of navigation points in the graph, rather than a single shortest path. This establishes a global path knowledge base in the storage module. Subsequently, it avoids the problem of getting trapped in local optima caused by each heavy-duty AGV independently using path planning algorithms such as A* algorithm and Dijkstra algorithm.

[0018] 2. The heavy-duty AGV transportation system based on central scheduling provided by this invention transfers all modules that may originally exist in the heavy-duty AGV, such as storage modules, task processing modules, and control modules, to the central scheduling system. This allows the heavy-duty AGV to serve only as the final execution unit for heavy-duty work, specializing in heavy-duty transportation, without the need for complex operations such as task splitting and path planning. This reduces weight while improving the stability and adaptability of the heavy-duty AGV to complex working conditions, reducing the probability of failures and conflicts. Furthermore, in application scenarios that require a large number of heavy-duty AGVs, the simplified structure of the AGVs also significantly reduces costs.

[0019] 3. The heavy-duty AGV transportation system based on central scheduling provided by this invention utilizes a unified central scheduling system to achieve collaborative scheduling and dynamic conflict resolution of several heavy-duty AGVs. When allocating tasks and selecting paths, the scheduling system can allocate appropriate tasks and non-conflicting or spatially and temporally staggered movement paths to multiple heavy-duty AGVs based on the global real-time status (such as the position of each heavy-duty AGV, task progress, and path occupancy). This effectively solves the deadlock and congestion problems that are easily caused by the lack of coordination among multiple AGVs at intersections and narrow passages in traditional distributed or independent planning modes. In particular, it meets the stringent requirements of high safety and high smoothness for heavy-duty AGV operations.

[0020] 4. The heavy-duty AGV transportation method based on central scheduling provided by this invention achieves a dual improvement in scheduling efficiency and system robustness by pre-computing a global path knowledge base and dynamically selecting and allocating heavy-duty AGVs. During the operation phase, this method avoids complex real-time full-network path search, and the scheduling decision response speed is extremely fast. At the same time, since multiple optimized paths are prepared for each movement, when the optimal path becomes unavailable due to unforeseen circumstances (such as temporary obstacles or other heavy-duty AGVs occupying the path), the system can immediately switch to the suboptimal alternative path without replanning, thereby ensuring the continuity and high reliability of the production process and significantly improving the overall logistics efficiency. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the heavy-duty AGV transportation system based on central scheduling according to the present invention. Figure 2 This is a flowchart of the scheduling method for the heavy-duty AGV transportation system based on central scheduling according to the present invention; The attached figures are labeled as follows: 1-Storage module, 2-Task processing module, 3-Control module, 4-Communication module, 5-Monitoring module, 6-Heavy-duty AGV, 7-Upper control system, 8-Central dispatching system. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0023] A heavy-duty AGV transportation system based on central scheduling, the schematic diagram of which can be found here. Figure 1 It includes a central dispatch system that communicates with the upper control system, and 100 heavy-duty AGVs that are uniformly dispatched by the central dispatch system.

[0024] The central dispatch system includes a storage module, a task processing module, a control module, a communication module, and a monitoring module.

[0025] The storage module stores the device action information database and the global path knowledge base. The action information database contains control parameters corresponding to all controllable atomic actions of the heavy-duty AGV; the global path knowledge base contains path information pre-calculated based on navigation points on the work area map. The device action information database is stored in a horizontally partitioned database format, including a point table, a line table, a device information table, and a neighboring point relationship table. The point table stores the static attributes of each navigation point on the map, including point number, X-coordinate, Y-coordinate, orientation angle, and point status. The line table stores the attributes of the path segment connecting two navigation points, including path segment number, start number, end number, path length, preset empty speed, preset loaded speed, and path direction. The device information table stores the static attributes and dynamic status of each heavy-duty AGV; static attributes include device number, device name, MAC address, load capacity, size information, speed, and acceleration. The neighboring point relationship table stores pairs of navigation points with direct connectivity on the work area map to represent the topology of the path network. In actual application scenarios, the items and data structure in the device action information database can also be adjusted according to actual needs.

[0026] The task processing module communicates with the upper control system to receive and parse the work tasks from the upper control system, break down the work tasks into an ordered sequence of atomic actions, and return the work task execution progress status information to the upper control system.

[0027] The control module is connected to the storage module and the task processing module respectively. It is used to issue control commands to the heavy-duty AGV based on the device action information database and global path knowledge database in the storage module, as well as the ordered atomic action sequence obtained by the task processing module.

[0028] The communication module connects to the control module and to each heavy-duty AGV, used to send control commands to designated heavy-duty AGVs and receive status signals from each heavy-duty AGV. The data communication protocol between the communication module and each heavy-duty AGV is Modbus (in practical applications, TCP, Socket, etc., can also be used), and the polling period is less than or equal to 100ms. Control commands include movement control commands and device operation control commands. For atomic actions involving movement, the control module selects a path for the designated heavy-duty AGV from the global path knowledge base and generates movement control commands by combining the corresponding control parameters queried from the device action information base. For atomic actions not involving movement, the control module queries the corresponding control parameters from the device action information base and generates device operation control commands.

[0029] The monitoring module communicates with the communication module to monitor the equipment status information, real-time position and motion status information, and task execution progress status information of the heavy-duty AGV based on status signals. The equipment status information includes battery level, battery health, voltage and current during charging, and equipment fault codes; the real-time position and motion status information includes X-axis coordinates, Y-axis coordinates, real-time speed, heading angle, and motion control mode; the task execution progress status information includes the currently executing task number, the atomic action instructions being executed under this task, and the completion status of the atomic actions.

[0030] This embodiment also provides a scheduling method for the aforementioned heavy-duty AGV transportation system based on central scheduling; see flowchart below. Figure 2 Its characteristic is that it includes the following steps: Step 1: Build a heavy-duty AGV transportation system based on central dispatch, set up work areas and establish a global path knowledge base for these areas; Step 1.1: Define the working area. Based on the line table and nearest-neighbor relationship table in the equipment motion information database, establish a directed weighted graph G={V,E,W}, where V is the vertex set, E is the edge set, and W is the weight matrix; V={v1,v2…v n}, where n is the number of navigation points in the working area, v1, v2…v n Representing different navigation points; for any edge e in the edge set E, represented by the triple (v i ,v j ,w ij ) indicates that v i and v j Let w represent the starting and ending navigation points of edge e, respectively. ij This indicates that the heavy-load AGV starts from the initial navigation point v. i Drive directly to the final navigation point v j The cost of passage; the w ij The elements that constitute the corresponding positions in the weight matrix W; if there is no direct path between two navigation points, then w is defined. ij =∞, and the travel cost w from a navigation point to itself ii =0; Step 1.2: Initialize an n×n travel cost matrix dist. Let the initial travel cost matrix dist... (0) =Weight matrix W, defined dist (0) [i][j]=w ij The initial passage cost matrix dist (0) As the initial state for iteration; Step 1.3: Sequentially select each navigation point in the vertex set V as a candidate transit navigation point v. kFor the initial passage cost matrix dist (0) Begin iterative updates, where k = 1, 2, ..., n; During each iteration, for each navigation point pair (v) i ,v j Perform the following operations: a. Calculation toll costs ,in This is the toll cost matrix after the previous iteration; b. Based on and The corresponding path determines the starting navigation point v. i transit navigation point v k To the final navigation point v j Complete vertex sequence ; c. Path information tuple Add to navigation point pair (v i ,v j The corresponding set of alternative paths middle; This embodiment uses an optimized variant of the classic Floyd-Warshall algorithm. The classic Floyd-Warshall algorithm is a dynamic programming algorithm for finding the shortest path from all sources. Its design purpose and output result is to determine a unique shortest path (i.e., the path with the lowest global cost) for each pair of vertices in a directed weighted graph.

[0031] In contrast, the core design objective of the optimized variant algorithm described in this embodiment is to not only track and update the shortest path between any two points during the algorithm iteration process, but also to systematically record and maintain a set of alternative paths containing multiple better paths. Its direct output is to provide multiple optimized path options (such as the first five) sorted by cost for each pair of connectable vertices in the graph, rather than a single solution. This fundamental difference gives the scheduling system of this embodiment essentially different technical characteristics and effects. The single deterministic path provided by classic algorithms is prone to task interruption or waiting in real-world dynamic industrial environments due to temporary path occupancy, blockage, or equipment malfunction. This embodiment, by providing multiple pre-computed optimized alternative paths, enables the central scheduling system to dynamically select the most suitable available path based on the global real-time status (such as path occupancy and equipment priority) during runtime. This endows the system with the inherent ability to cope with dynamic interference, make online decisions, and replan, thereby significantly enhancing the robustness, flexibility, and overall efficiency of the entire heavy-duty AGV fleet operation, achieving a technological leap from static optimal planning to dynamic robust scheduling.

[0032] Step 1.4: After completing all iterations, based on each navigation point pair (v) i ,v j ), and its set of alternative paths All paths are sorted in ascending order based on their travel costs; the top five paths with the lowest travel costs are selected and stored in the storage module as a global path knowledge base; in this embodiment, the travel cost is distance. In actual application scenarios, multiple factors such as factory layout, peak and off-peak electricity prices, number of curves in the path, safety risk factors, and other weighted factors can also be considered.

[0033] Step 2: Deploy 100 of the heavy-duty AGVs into the work area; Step 3: Start working. The upper control system sends the task to the central dispatch system. The task processing module of the central dispatch system breaks down the task into an ordered sequence of atomic actions. Step 4: The control module iterates through all heavy-load AGVs via the communication module and selects and designates one or more heavy-load AGVs that can execute the above atomic action sequence; Step 5: For each atomic action in the atomic action sequence, query the corresponding control parameters from the device action information database; if the atomic action involves movement, perform path planning for each specified heavy-duty AGV from the global path knowledge base established in Step 1, specifically: Step a1: Based on the starting point and ending point of the current atomic action, retrieve five pre-stored alternative paths and their distances from the global path knowledge base for the starting point and ending point pair; Step a2: Check the real-time occupancy status of the five candidate paths, and select the path with the shortest distance among the unoccupied paths as the execution path; Then, based on the retrieved control parameters and the planned path, control commands are generated.

[0034] Step 6: Send the control command to each designated heavy-duty AGV to drive it to execute the atomic action sequence in sequence. During and after execution, the task processing module returns the task execution progress status information to the upper control system.

Claims

1. A heavy-duty AGV transportation system based on central scheduling, characterized in that: This includes a central dispatching system that communicates with the upper control system, and several heavy-duty AGVs that are uniformly dispatched by the central dispatching system; The central dispatch system includes a storage module, a task processing module, a control module, a communication module, and a monitoring module. The storage module is used to store a device action information database and a global path knowledge base. The action information database contains control parameters corresponding to all controllable atomic actions of the heavy-load AGV; the global path knowledge base contains path information pre-calculated based on navigation points on the work area map. The task processing module is communicatively connected to the upper control system and is used to receive and parse the work tasks from the upper control system, break down the work tasks into an ordered sequence of atomic actions, and return the work task execution progress status information to the upper control system. The control module is connected to the storage module and the task processing module respectively, and is used to issue control commands to the specified heavy-duty AGV based on the device action information database and global path knowledge database in the storage module, and the ordered atomic action sequence obtained by the task processing module. The communication module is connected to the control module and to each heavy-duty AGV, and is used to send the control commands to the specified heavy-duty AGV and receive the status signals returned by each heavy-duty AGV. The monitoring module is connected to the communication module and is used to monitor the equipment body status information, real-time position and motion status information, and task execution progress status information of the heavy-duty AGV according to the status signal.

2. The heavy-duty AGV transportation system based on central dispatch according to claim 1, characterized in that: The equipment action information database is stored in a horizontally partitioned database format, which includes a point table, a line table, an equipment information table, and a neighboring point relationship table. The point table is used to store the static attributes of each navigation point in the map, including point number, X-coordinate, Y-coordinate, orientation angle, and point status; The line table is used to store the attributes of the path segment connecting two navigation points, including the path segment number, start number, end number, path length, preset empty speed, preset loaded speed, and path direction attribute. The equipment information table stores the static attributes and dynamic status of each heavy-duty AGV; the static attributes include equipment number, equipment name, MAC address, load capacity, size information, speed and acceleration; The nearest point relationship table is used to store pairs of navigation points with direct connectivity in the working area map to represent the topology of the path network.

3. The heavy-duty AGV transportation system based on central dispatch according to claim 2, characterized in that: The control instructions include movement control instructions and equipment operation control instructions. For atomic actions involving movement, the control module selects a path for the specified heavy-duty AGV from the global path knowledge base and generates movement control instructions by combining the corresponding control parameters queried from the equipment action information base. For atomic actions not involving movement, the control module queries the corresponding control parameters from the equipment action information base and generates equipment operation control instructions.

4. The heavy-duty AGV transportation system based on central dispatch according to claim 3, characterized in that: The data communication protocol between the communication module and each heavy-duty AGV is the Modbus protocol, and the polling period is less than or equal to 100ms.

5. The heavy-duty AGV transportation system based on central dispatch according to claim 4, characterized in that: The device status information includes battery power, battery health, voltage and current during charging, and device fault codes. The real-time position and motion status information includes X-axis coordinates, Y-axis coordinates, real-time speed, heading angle, and motion control mode; The task execution progress status information includes the currently executing task number, the atomic action instructions being executed under the task, and the completion status of the atomic actions.

6. A scheduling method for a heavy-duty AGV transportation system based on central scheduling as described in any one of claims 1 to 5, characterized in that, Includes the following steps: Step 1: Build a heavy-duty AGV transportation system based on central dispatch, set up work areas and establish a global path knowledge base for these areas; Step 2: Deploy several heavy-duty AGVs into the work area; Step 3: Start working. The upper control system sends the task to the central dispatch system. The task processing module of the central dispatch system breaks down the task into an ordered sequence of atomic actions. Step 4: The control module traverses all heavy-load AGVs through the communication module and selects and designates one or more heavy-load AGVs that can execute the above atomic action sequence; Step 5: For each atomic action in the atomic action sequence, query the corresponding control parameters from the device action information database; if the atomic action involves movement, perform path planning for each specified heavy-duty AGV from the global path knowledge base established in Step 1; generate control instructions based on the queried control parameters and the planned path. Step 6: Send the control command to each designated heavy-duty AGV to drive it to execute the atomic action sequence in sequence. During and after execution, the task processing module returns the task execution progress status information to the upper control system.

7. The scheduling method for a heavy-duty AGV transportation system based on central scheduling according to claim 6, characterized in that: Step 1 is as follows: Step 1.1: Define the working area. Based on the line table and nearest-neighbor relationship table in the equipment motion information database, establish a directed weighted graph G={V,E,W}, where V is the vertex set, E is the edge set, and W is the weight matrix; V={v1,v2…v n }, where n is the number of navigation points in the working area, v1, v2…v n Representing different navigation points; for any edge e in the edge set E, represented by the triple (v i ,v j ,w ij ) indicates that v i and v j Let w represent the starting and ending navigation points of edge e, respectively. ij This indicates that the heavy-load AGV starts from the initial navigation point v. i Drive directly to the final navigation point v j The cost of passage; the w ij The elements that constitute the corresponding positions in the weight matrix W; If there is no direct path between two navigation points, then define w. ij =∞, and the travel cost w from a navigation point to itself ii =0; Step 1.2: Initialize an n×n travel cost matrix dist. Let the initial travel cost matrix dist... (0) =Weight matrix W, defined dist (0) [i][j]=w ij The initial passage cost matrix dist (0) As the initial state for iteration; Step 1.3: Sequentially select each navigation point in the vertex set V as a candidate transit navigation point v. k For the initial passage cost matrix dist (0) Begin the iteration, where k = 1, 2, ..., n; During each iteration, for each navigation point pair (v) i ,v j Perform the following operations: a. Calculation toll costs ,in This is the toll cost matrix after the previous iteration; b. Based on and The corresponding path determines the starting navigation point v. i transit navigation point v k To the final navigation point v j Complete vertex sequence ; c. Path information tuple Add to navigation point pair (v i ,v j The corresponding set of alternative paths middle; Step 1.4: After completing all iterations, based on each navigation point pair (v) i ,v j ), and its set of alternative paths All paths are sorted in ascending order based on their travel cost; the top five paths with the lowest travel cost are selected and stored in the storage module as a global path knowledge base.

8. The scheduling method for a heavy-duty AGV transportation system based on central scheduling according to claim 7, characterized in that: In steps 1.1-1.4, the travel cost is distance.

9. The scheduling method for a heavy-duty AGV transportation system based on central scheduling according to claim 8, characterized in that: In step 5, path planning is performed for each specified heavy-duty AGV from the global path knowledge base, specifically as follows: a1. Based on the starting point and ending point of the current atomic action, obtain from the global path knowledge base five pre-stored alternative paths and their distances for the navigation point pair formed by the starting point and ending point. a2. Check the real-time occupancy status of the five candidate paths, and select the path with the shortest distance among the unoccupied paths as the execution path.

Images