Multi-layer storage multi-vehicle-type cooperative scheduling method and system

Abstract

The invention discloses a multi-layer storage multi-vehicle-type collaborative scheduling system and method, and belongs to the technical field of intelligent storage logistics, and the method comprises the steps that a warehouse management system issues a material carrying task to a robot control system; the robot control system selects an optimal junction point; the robot control system dispatches the stacker to go to a task starting point to pick goods and send the goods to a handover point; the robot control system dispatches the AGV to receive the cargoes and convey the cargoes to a task end point at the same time; and the robot control system feeds back task completion information and releases double-vehicle resources. According to the invention, through dynamic handover point selection, double-vehicle synchronous scheduling and handover area control, the labor cost is reduced, the problems of low multi-vehicle cooperation efficiency, long waiting time and collision risk are solved, and full-automatic safe operation is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to a multi-layer warehouse multi-vehicle type collaborative scheduling method and system, belonging to the technical field of intelligent warehouse logistics. BACKGROUND

[0002] Under the background of logistics automation upgrading, both stacker trucks (unmanned forklifts) and AGVs (Automated Guided Vehicles, a kind of unmanned intelligent handling equipment) are needed to work collaboratively in a multi-layer warehouse system. Since stacker trucks are responsible for vertical access to high shelves, and AGVs are responsible for horizontal transfer on the ground, the two types of equipment have different functions and independent scheduling logic, resulting in low collaborative efficiency of multi-vehicle type, which has become a key bottleneck restricting the improvement of warehouse automation level.

[0003] The existing method limits the division of labor of equipment by defining exclusive work areas, and relies on manual triggering or fixed signal equipment to link double-car tasks. This mode has the following fundamental defects: manual intervention leads to response delay, and there is a risk of false triggering; fixed handover points cause equipment congestion, and the collision probability increases when multiple tasks are parallel. For example, the existing work process has significant dependence on manual intervention: when the stacker truck arrives at the fixed handover point with the goods, it needs to be manually scanned and confirmed (taking 10-15 seconds), and then the instruction is triggered to call the AGV; after receiving the instruction, the AGV drives to the handover point to perform the lifting and transfer. The actual measurement data shows that the waiting interval from the completion of the stacker truck unloading to the start of the AGV operation is 30 seconds to 2 minutes, and this delay increases with the expansion of the warehouse area.

[0004] Therefore, how to improve the multi-layer warehouse multi-vehicle type collaborative scheduling method and system to realize dynamic binding of double-car tasks, real-time optimization of handover points, and collaborative control of paths, so as to eliminate manual intervention and improve the efficiency of equipment collaboration, is a technical problem to be solved by those skilled in the art. SUMMARY

[0005] In view of the above deficiencies in the prior art, the purpose of the present application is to provide a multi-layer warehouse multi-vehicle type collaborative scheduling method and system to solve the problems of low work efficiency, fixed handover points causing equipment congestion, and delayed task connection in the existing solutions.

[0006] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0007] A multi-layer warehouse multi-vehicle type collaborative scheduling method, characterized in that it comprises the following steps:

[0008] S10, the warehouse management system issues a material handling task to the robot control system, and the robot control system determines whether double-car collaboration is needed according to whether the task starting point and the task ending point are located in different work areas;

[0009] S20, the robot control system selects an optimal handover point from the set of idle handover points based on the task start point and end point positions by a calculation formula;

[0010] S30, the robot control system uses a greedy algorithm to schedule the idle stacker closest to the task start point, controls the stacker to go to the task start point according to the generated path to pick up goods and deliver them to the handover point;

[0011] S40, the robot control system simultaneously uses a greedy algorithm to schedule the idle AGV closest to the handover point, controls the AGV to receive goods according to the generated path and deliver them to the task end point;

[0012] S50, the robot control system feeds back the overall task completion information to the warehouse management system after receiving the stacker unloading completion state and the AGV delivery completion state, and releases the double-car resources to the idle queue.

[0013] Further, the S20 specifically comprises the following steps:

[0014] S21, the robot control system obtains the current set of idle handover points {P1, P2,..., P k};

[0015] S22, the robot control system obtains the task start point coordinates (x s , y s ), the task end point coordinates (x e , y e ), and the i-th handover point coordinates (x i , y i ) according to the material handling task information, and calculates the formula:

[0016] D i = |x s -x i | + |y s -y i | + |x i -x e | + |y i -y e |;

[0017] S23, according to the calculated D i , select the handover point P i with the smallest D * as the handover point of this task.

[0018] Further, the S30 specifically comprises the following steps:

[0019] S31, the robot control system obtains the set of idle stackers as {F1, F2,..., F m, its current position is The task starting point is (x s ,y s ), by calculating and comparing the Manhattan distances of all idle stackers to the task starting point, the stacker F j* that has the smallest distance is selected to perform the task; the selection formula is as follows:

[0020]

[0021] Wherein, j * is the index number of the selected stacker in the set;

[0022] S32, the path planning unit of the robot control system uses an improved A* algorithm to generate the following two paths in turn:

[0023] The first path is empty driving to pick up goods: generate a driving path Path1 from the current position N start of the selected stacker F j to the task starting point S=(x s ,y s );

[0024] The second path is carrying goods to deliver goods: generate a driving path Path2 from the task starting point S=(x s ,y s ) to the optimal handover point P=(x * ,y * );

[0025] The specific execution steps of the improved A* algorithm are as follows:

[0026] Input: warehouse grid map (where the obstacle node value is 1 and the free passage node value is 0), starting node coordinates (N start for Path1 and S for Path2), target node coordinates (S for Path1 and P for Path2),

[0027] Node evaluation function: for any node n to be evaluated, its evaluation function f(n) is calculated by the following formula:

[0028] f(n) = g(n) + h(n)

[0029] Wherein, g(n) represents the actual movement cost from the starting node to node n, taking the actual number of grids passed, h(n) is the heuristic function from node n to the target node, calculated using Manhattan distance:

[0030] h(n) = |x n -x g | + |y n -y g |

[0031] where (x g ,y g ) is the coordinate of the target node;

[0032] Improved heuristic function (for improving multi-vehicle coordination efficiency):

[0033] To further optimize multi-vehicle path conflict, dynamic conflict cost is introduced, and the heuristic function can be extended as:

[0034] h(n) = d goal (n) + a·ConflictCost(n)

[0035] where d goal (n) is the geometric distance (Manhattan distance) from node n to the target node, ConflictCost(n) is the conflict cost dynamically calculated according to historical path data, which is provided by the communication and coordination module, representing the expected number of conflicts within the future time window, and a is the conflict cost coefficient, which is empirically valued between [0.1, 0.5];

[0036] Algorithm execution: use open list and closed list to traverse nodes, select the node with the smallest f(n) value each time for expansion, until the target node P * is reached.

[0037] Output: generate the globally optimal path node sequence Path = [N1, N2,..., N k ], and the path planning unit serializes the path instructions Path1 and Path2 and sends them to the stacker control module in turn;

[0038] Step S33, the stacker control module receives the stacker control instruction and drives the stacker to first move along Path1 to the task starting point to pick up the goods, and then move along Path2 to the handover point to unload the goods;

[0039] Step S34, when the stacker unloads the goods to the handover point, the stacker control module generates a unloading completion instruction and reports it to the robot control system.

[0040] Further, the S40 specifically includes the following steps:

[0041] S41, the robot control system obtains the idle AGV set as {A1, A2,..., A n}, where the current position coordinates of AGVA k are The optimal handover point coordinates of this task are (x * , y *), the AGV with the smallest Manhattan distance to the handover point is selected by calculating and comparing the Manhattan distances of all idle AGVs to the handover point; the selection formula is as follows:

[0042]

[0043] wherein, k * is the index number of the selected AGV in the set;

[0044] S42, the path planning unit of the robot control system generates the travel path Path3 from the current position of the selected AGV to the optimal handover point P=(x * ,y * ) by using the same improved A* algorithm as S32 and delivers the path to the AGV control module;

[0045] S43, the AGV control module receives the AGV control instruction and drives the AGV to reach a preset point outside the handover area to wait;

[0046] S44, the AGV waits at the waiting point and continuously reports the position state;

[0047] S45, when the robot control system receives the forklift unloading completion state and confirms that the handover area is idle, the second stage access instruction is delivered to the AGV control module;

[0048] S46, the AGV control module drives the AGV to enter the handover area to pick up goods, and according to the travel path Path4 from the optimal handover point P=(x * ,y * ) to the task end point G=(x e ,y e ) generated by the improved A* algorithm, the goods are transported to the task end point and the task completion state is reported to the robot control system.

[0049] The application also provides a multi-layer warehouse multi-vehicle type collaborative scheduling system, comprising: a warehouse management system, a robot control system and a map server; at least one forklift, each forklift comprising a forklift body and a forklift control module connected thereto; at least one AGV, each AGV comprising an AGV body and an AGV control module connected thereto; the warehouse management system is in communication connection with the robot control system; the map server is connected with the robot control system; the forklift control module and the AGV control module are connected with the robot control system respectively;

[0050] The warehouse management system and the robot control system perform the multi-layer warehouse multi-vehicle type collaborative scheduling method of the application, comprising:

[0051] The warehouse management system is used to deliver the material handling task to the robot control system;

[0052] The robot control system is used to analyze task requirements and confirm that the stacker truck and AGV need to work together. Based on Manhattan distance calculation, it selects the optimal handover point from the set of idle handover points, generates the driving path of the stacker truck and AGV, and coordinates the task sequence and handover area access of the stacker truck and AGV.

[0053] The map server provides warehouse map data and the basis for Manhattan distance calculation;

[0054] The stacker control module receives path instructions from the robot control system;

[0055] The AGV control module receives two-stage control commands (waiting command + admission command) from the robot control system.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. This invention, through methodological and system innovation, solves the problems of low operational efficiency, equipment congestion caused by fixed handover points, and untimely task coordination in existing solutions. Furthermore, by employing a dynamic handover point selection mechanism and a dual-vehicle synchronous scheduling strategy, it achieves efficient collaborative operation between stacker trucks and AGVs within a unified map framework, completely eliminating manual scanning and related job requirements, and significantly reducing labor costs.

[0058] 2. This invention dynamically selects the optimal handover point based on the Manhattan distance model, effectively shortening the equipment transfer path. By scheduling stacker trucks and AGVs in parallel, it transforms the traditional serial task process into parallel operation, significantly reducing or even eliminating task connection waiting time. This solution fills the technical gap in fully automated collaborative scheduling of multiple vehicle types in multi-layer warehousing scenarios, significantly improving equipment utilization and overall warehousing operation efficiency while ensuring operational safety and system compatibility.

[0059] 3. This invention is based on unmanned forklifts (forklift trucks) and AGVs. When these two types of equipment work together, strict safety regulations must be met: after unloading, the forks of the forklift must be fully extended and the forklift placed stably; the movement of the AGV is constrained by the accessibility of the path. The forklift has the capability to store and retrieve goods at heights of 3-10 meters; the AGV is a stealthy AGV, using QR code navigation, with a lifting height of 60mm, a load capacity of 1000KG, a rated unloaded speed of 1.8m / s, a full-load speed of 1.5m / s, a positional accuracy of ±10mm, and is equipped with laser obstacle avoidance and collision detection, increasing the AGV's adaptive obstacle avoidance capabilities. Attached Figure Description

[0060] Figure 1 This is a flowchart of the multi-layer warehouse and multi-vehicle collaborative scheduling method of Embodiment 1 of the present invention;

[0061] Figure 2 is a block diagram of a multi-layer warehouse multi-vehicle cooperative scheduling system according to an embodiment of the present application;

[0062] Figure 3 is a schematic diagram of a computer system for implementing the scheduling method. DETAILED DESCRIPTION

[0063] The technical solutions of the present application will be described in detail below with reference to the embodiments and drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0064] Embodiment 1

[0065] As shown in Figure 1 , the present application provides a multi-layer warehouse multi-vehicle cooperative scheduling method, comprising the following steps:

[0066] S10, the warehouse management system issues a material handling task to the robot control system, and the robot control system determines whether double-vehicle cooperation is required according to whether the task starting point and the task ending point are located in different work areas (such as high-rack area and ground transfer area);

[0067] S20, the robot control system selects an optimal handover point from the set of idle handover points based on the task starting point position and the task ending point position by using a calculation formula;

[0068] S30, the robot control system uses a greedy algorithm to schedule an idle reach truck closest to the task starting point, and controls the reach truck to take goods from the task starting point to the handover point according to the generated path;

[0069] S40, the robot control system simultaneously uses a greedy algorithm to schedule an idle AGV closest to the handover point, and controls the AGV to receive goods and deliver them to the task ending point according to the generated path;

[0070] S50, the robot control system feeds back the overall task completion information to the warehouse management system after receiving the reach truck unloading completion state and the AGV delivery completion state, and releases the double-vehicle resources to the idle queue.

[0071] S20 specifically comprises the following steps:

[0072] S21, the robot control system obtains the current set of idle handover points {P1, P2,..., P k};

[0073] S22, the robot control system obtains the task starting point coordinates (x s,y s The endpoint coordinates of the task are (x e ,y e The coordinates of the i-th intersection point are (x... i ,y i ), Calculation formula:

[0074] D x =|x s -x i |+|y s -y i |+|x i -x e |+|y i -y e |;

[0075] S23, based on the calculated D i Choose D i Minimum intersection point P * This serves as the handover point for this mission.

[0076] S30 includes the following steps:

[0077] S31, the robot control system obtains the set of idle stacker trucks as {F1, F2, ..., F...} m}, its current position is The starting point of the task is (x s ,y s By calculating and comparing the Manhattan distances of all idle forklift trucks to the task starting point, the forklift truck F with the smallest distance is selected. j* Execute the task; the selection formula is as follows:

[0078]

[0079] Where, j * That is, the index number of the selected forklift truck in the set;

[0080] S32, the path planning unit of the robot control system uses an improved A* algorithm to generate the following two path segments sequentially:

[0081] First route (empty pickup): Generates a pickup route from the selected stacker truck F. j Current position N start To the starting point of the task S = (x s ,y s The driving path Path1;

[0082] The second path (cargo delivery): generates the path from the task starting point S = (x s ,y s ) to the optimal intersection point P = (x *y * ) of the driving path Path2;

[0083] The specific execution steps of the improved A* algorithm are as follows:

[0084] Input: warehouse grid map (where the obstacle node value is 1 and the free passage node value is 0), starting node coordinates (N start for Path1 and S for Path2), target node coordinates (S for Path1 and P for Path2),

[0085] Node evaluation function: for any node n to be evaluated, its evaluation function f(n) is calculated by the following formula:

[0086] f(n) = g(n) + h(n)

[0087] where g(n) represents the actual movement cost from the starting node to node n, taking the actual number of grids passed, h(n) is the heuristic function from node n to the target node, calculated using Manhattan distance:

[0088] h(n) = |x n -x g | + |y n -y g |

[0089] where (x g , y g ) is the coordinate of the target node;

[0090] Improved heuristic function (for improving multi-vehicle coordination efficiency):

[0091] To further optimize multi-vehicle path conflict, a dynamic conflict cost is introduced, and the heuristic function can be extended to:

[0092] h(n) = d goal (n) + α·ConflictCost(n)

[0093] where d goal (n) is the geometric distance (Manhattan distance) from node n to the target node, ConflictCost(n) is the conflict cost dynamically calculated according to historical path data, which is provided by the communication coordination module and represents the expected number of conflicts within a future time window, and α is the conflict cost coefficient, which is empirically valued between [0.1, 0.5];

[0094] Algorithm execution: use open list and closed list to traverse nodes, select the node with the smallest f(n) value each time for expansion, until the target node P * is reached,

[0095] Output: the global optimal path node sequence Path = [N1, N2,..., N k ] is generated, and the path planning unit sequentially issues the serialized path instructions Path1 and Path2 to the stacker control module.

[0096] Step S33, the stacker control module receives the stacker control instruction and drives the stacker to first move along Path1 to the task starting point to pick up goods, and then move along Path2 to the handover point to unload the goods;

[0097] Step S34, when the stacker unloads the goods to the handover point, the stacker control module generates an unloading completion instruction and reports it to the robot control system.

[0098] Wherein, the S40 specifically includes the following steps:

[0099] S41, the robot control system obtains the idle AGV set as {A1, A2,..., A n}, wherein the current position coordinates of AGVA k The optimal handover point coordinates of this task are (x * , y * ), by calculating and comparing the Manhattan distances of all idle AGVs to the handover point, the AGV with the smallest distance is selected to perform the task; the selection formula is as follows:

[0100]

[0101] Wherein, k * is the index number of the selected AGV in the set;

[0102] S42, the path planning unit of the robot control system generates the driving path Path3 from the current position of the selected to the optimal handover point P = (x * , y * ) using the same improved A* algorithm as S32 and issues it to the AGV control module;

[0103] S43, the AGV control module receives the AGV control instruction and drives the AGV to reach the preset point outside the handover area to wait;

[0104] S44, the AGV waits at the waiting point and continuously reports the position state;

[0105] S45, when the robot control system receives the stacker unloading completion state and confirms that the handover area is idle, the second stage access instruction is issued to the AGV control module;

[0106] ​S46, the AGV control module drives the AGV to enter the handover area to pick up the goods, and generates a driving path Path4 from the optimal handover point P=(x * ,y * ) to the task end point G=(x e ,y e ) according to the improved A* algorithm, and transports the goods to the task end point and reports the task completion status to the robot control system.

[0107] The present application realizes efficient collaborative operation of the forklift and the AGV in the unified map framework through the dynamic handover point selection mechanism and the double-vehicle synchronous scheduling strategy. The artificial scanning link and related post requirements are completely eliminated, and the labor cost is significantly reduced. At the same time, the optimal handover point is dynamically selected based on the Manhattan distance model, and the equipment transfer path is effectively shortened. The traditional serial task process is changed into parallel operation through parallel scheduling of the forklift and the AGV, and the task connection waiting time is greatly compressed or even eliminated. The present application fills the technical gap of multi-vehicle type full-automatic collaborative scheduling in the multi-layer warehouse scenario, while ensuring the operation safety and system compatibility, significantly improves the comprehensive utilization rate of equipment and the overall warehouse operation efficiency.

[0108] Referring to Figure 2 , the multi-layer warehouse multi-vehicle type collaborative scheduling method and the multi-layer warehouse multi-vehicle type collaborative scheduling system are implemented. The multi-layer warehouse multi-vehicle type collaborative scheduling system comprises: a warehouse management system 201, a robot control system 202, a map server 203; at least one forklift 204, each forklift comprising a forklift body and a forklift control module 204a connected thereto; at least one AGV 205, each AGV comprising an AGV body and an AGV control module 205a connected thereto; the warehouse management system 201 and the robot control system 202 are in communication connection; the map server 203 and the robot control system 202 are connected through a gigabit Ethernet; the forklift control module 204a and the AGV control module 205a are respectively connected with the robot control system 202 through a 5G industrial router;

[0109] The warehouse management system 201 is used to issue material handling tasks to the robot control system 202;

[0110] The robot control system 202 comprises: a task analysis module for analyzing task requirements and confirming that the forklift 204 and the AGV 205 need to be cooperatively executed, a dynamic handover point selection module for selecting an optimal handover point from a set of idle handover points based on Manhattan distance calculation, a path planning unit for generating driving paths of the forklift 204 and the AGV 205, and a double-vehicle synchronous scheduling module for coordinating task timing and handover area access of the forklift 204 and the AGV 205;

[0111] The map server 203 provides warehouse map data and Manhattan distance calculation basis;

[0112] The stacker control module 204a receives path instructions of the robot control system 202.

[0113] The AGV control module 205a receives two-stage control instructions (waiting instruction + access instruction) of the robot control system 202.

[0114] Further, referring to Figure 3 Also provided is a non-transitory computer readable storage medium including instructions, such as a memory 302 including instructions, which can be executed by the processor 301 of the above-mentioned device to complete the multi-layer warehouse multi-vehicle collaborative scheduling method. For example, the non-transitory computer readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0115] In an exemplary embodiment, an application product is also provided, including one or more instructions which can be executed by the processor 301 of the above-mentioned device to complete the above-mentioned multi-layer warehouse multi-vehicle collaborative scheduling method.

[0116] Note that the above is only a preferred embodiment of the present application and the technical principles applied. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and those skilled in the art can make various obvious changes, readjustments and substitutions without departing from the scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A multi-layer warehouse and multi-vehicle collaborative scheduling method, characterized in that, Includes the following steps: S10, the warehouse management system sends material handling tasks to the robot control system. The robot control system determines whether two robots need to work together based on whether the task start and end points are located in different work areas. S20, the robot control system selects the optimal handover point from the set of idle handover points based on the task start position and end position using a calculation formula; S30, the robot control system uses a greedy algorithm to schedule the idle stacker truck closest to the task start point, and controls the stacker truck to pick up the goods from the task start point and deliver them to the handover point according to the generated path. S40, the robot control system simultaneously uses a greedy algorithm to schedule the idle AGV closest to the handover point, and controls the AGV to receive the goods and transport them to the task endpoint according to the generated path. S50, after receiving the unloading completion status of the stacker truck and the delivery completion status of the AGV, the robot control system feeds back the overall task completion information to the warehouse management system and releases the dual-vehicle resources to the idle queue.

2. The multi-layer warehouse and multi-vehicle collaborative scheduling method according to claim 1, characterized in that, S20 specifically includes the following steps: S21, the robot control system obtains the current set of idle handover points {P1, P2, ..., P...} k }; S22, the robot control system obtains the task starting coordinates (x) based on the material handling task information. s ,y s The endpoint coordinates of the task are (x e ,y e The coordinates of the i-th intersection point are (x... i ,y i ), Calculation formula: D i =|x s -x i |+|and s -and i |+|x i -x e |+|and i -and e |; S23, based on the calculated D i Choose D i Minimum intersection point P * This serves as the handover point for this mission.

3. The multi-layer warehouse and multi-vehicle collaborative scheduling method according to claim 1, characterized in that, S30 specifically includes the following steps: S31, the robot control system obtains the set of idle stacker trucks as {F1, F2, ..., F...} m }, its current position is The starting point of the task is (x s ,y s The algorithm calculates and compares the Manhattan distances of all available forklift trucks to the task starting point, and selects the forklift truck with the smallest distance. Execute the task; the selection formula is as follows: Where, j * That is, the index number of the selected forklift truck in the set; S32, the path planning unit of the robot control system uses an improved A* algorithm to generate the following two path segments sequentially: The first segment of the route is empty pickup: generating a pickup from the selected stacker truck F. j Current position N start To the starting point of the task S = (x s ,y s The driving path Path1; The second path is for cargo delivery: This is generated from the task starting point S = (x s ,y s ) to the optimal intersection point P = (x * ,y * The driving path Path2; The specific execution steps of the improved A* algorithm are as follows: Input: A warehouse grid map, where obstacle nodes have a value of 1 and free passage nodes have a value of 0; the starting node coordinates, which are N for Path1. start For Path2, the coordinates are S; for the target node, the coordinates are S for Path1 and P for Path2. Node evaluation function: For any node n to be evaluated, its evaluation function f(n) is calculated by the following formula: f(n) = g(n) + h(n) Where g(n) represents the actual movement cost from the starting node to node n, taking the actual number of grid cells traversed, and h(n) is the heuristic function from node n to the target node, calculated using Manhattan distance: h(n)=|x n -x g |+|y n -y g | Among them, (x g ,y g ) are the coordinates of the target node; An improved heuristic function for enhancing multi-vehicle collaboration efficiency: To further optimize multi-vehicle path conflicts, a dynamic conflict cost is introduced, and the heuristic function can be extended as follows: h(n)=d goal (n)+α·ConflictCost(n) Where, d goal (n) is the geometric distance (Manhattan distance) from node n to the target node, ConflictCost(n) is the conflict cost dynamically calculated based on historical path data. This value is provided by the communication coordination module and represents the expected number of conflicts for the node in the future time window. α is the conflict cost coefficient, which is empirically taken between [0.1, 0.5]. Algorithm execution: Traverse the nodes using open and closed lists, selecting the node with the smallest f(n) value at each step for expansion, until the target node P is reached. * ; Output: Generate a globally optimal path node sequence Path = [N1, N2, ..., N k The path planning unit sequentially sends the serialized path instructions Path1 and Path2 to the stacker control module. Step S33: The stacker control module receives the stacker control command and drives the stacker to first move along Path1 to the task starting point to pick up the goods, and then move along Path2 to the handover point to unload the goods. Step S34: When the stacker truck unloads the goods to the handover point, the stacker truck control module generates an unloading completion instruction and reports it to the robot control system.

4. The multi-layer warehouse and multi-vehicle collaborative scheduling method according to claim 1, characterized in that, S40 specifically includes: S41, the robot control system obtains the set of idle AGVs as {A1, A2, ..., A...} n }, where AGVA k The current position coordinates are The optimal handover point coordinates for this task are (x * ,y * The selection process involves calculating and comparing the Manhattan distances from all idle AGVs to the handover point, and then selecting the AGV with the shortest distance to perform the task. The selection formula is as follows: Where, k * This is the index number of the selected AGV in the set; S42, the path planning unit of the robot control system uses the same improved A* algorithm as S32 to generate a path from the selected... The current position to the optimal intersection point P = (x * ,y * The driving path Path3 is then sent to the AGV control module. S43, the AGV control module receives the AGV control command and drives the AGV to a preset point outside the handover area to wait; S44, the AGV is in standby at the waiting point and continuously reports its position status; S45, when the robot control system receives the status of the stacker truck unloading completed and confirms that the handover area is empty, it sends the second-stage access instruction to the AGV control module; S46, the AGV control module drives the AGV into the handover area to retrieve goods, and according to the improved A* algorithm, retrieves goods from the optimal handover point P = (x * ,y * ) to the end point of the task G = (x e ,y e The robot will transport the goods to the destination along its Path4 route and report the task completion status to the robot control system.

5. A multi-layer warehouse and multi-vehicle collaborative scheduling system, characterized in that, include: The system includes a warehouse management system, a robot control system, and a map server; at least one stacker truck, each stacker truck including a stacker truck body and a stacker truck control module connected to it; at least one AGV, each AGV including an AGV body and an AGV control module connected to it; the warehouse management system is communicatively connected to the robot control system; the map server is connected to the robot control system; the stacker truck control module and the AGV control module are respectively connected to the robot control system. The warehouse management system and robot control system perform the method according to any one of claims 1 to 4: The warehouse management system is used to issue material handling tasks to the robot control system; The robot control system is used to analyze task requirements and confirm that the stacker truck and AGV need to work together. Based on Manhattan distance calculation, it selects the optimal handover point from the set of idle handover points, generates the driving path of the stacker truck and AGV, and coordinates the task sequence and handover area access of the stacker truck and AGV. The map server provides warehouse map data and the basis for Manhattan distance calculation; The stacker control module receives path instructions from the robot control system; The AGV control module receives two-stage control commands (waiting command + admission command) from the robot control system.

Images