AMR carrying robot dynamic path planning method

Abstract

The application discloses an AMR carrying robot dynamic path planning method. The method comprises the following steps: establishing a planning map based on a three-dimensional space-time grid matrix; receiving manufacturing execution service data and obtaining a target trolley unique identifier to extract physical three-dimensional geometric dimensions and load parameters; performing comprehensive cost evaluation calculation on candidate nodes, wherein the comprehensive cost function comprises a time cost, a service priority penalty coefficient and a mechanical envelope interference penalty; and generating a conflict-free scheduling path sequence based on the evaluation result and issuing the same. The application can realize production rhythm-based prediction planning to eliminate high-optimal task material blockage and completely eliminate the physical scratching risk and deadlock hidden danger of heterogeneous trolleys in a complex channel.

Description

Technical Field

[0001] This application relates to the field of data processing and management scheduling of computer flow networks, specifically a dynamic path planning method for AMR (Autonomous Mobile Robot) transport robots, which is particularly suitable for dynamic path scheduling of autonomous mobile robots based on manufacturing execution business data and trolley physical attributes in flexible logistics scenarios of discrete manufacturing. Background Technology

[0002] In modern high-end manufacturing, material flow typically relies on autonomous mobile robots towing or lifting various types of logistics trolleys between warehouses and production lines. Most existing autonomous mobile robot path planning systems depend on a single two-dimensional grid map, using conventional search algorithms to plan the shortest spatial distance from the starting point to the destination.

[0003] Currently, conventional technologies exhibit significant engineering shortcomings in real-world factory logistics management scenarios. Existing systems lack deep interaction with production data. When robots executing different order tasks interfere within limited aisle resources, the system only performs equidistant avoidance planning based on the physical distance detected by underlying sensors. It fails to recognize the priority differences in business logic between empty trolleys and trolleys carrying missing or stopped materials, easily leading to downtime on critical production lines due to delayed material delivery. Furthermore, existing planning algorithms treat robots and their loads as fixed-size point masses or standard circles. However, real factories contain various non-standard equipment such as material bin trolleys and long-axis hanging trolleys. The dynamic motion envelope of a robot changes significantly when combined with different trolleys. Fixed path planning models cannot adapt to the dynamic changes in trolley attributes, resulting in mechanical scraping or scheduling deadlocks between trolleys in densely populated work areas. How to eliminate the critical downtime risks caused by congestion and avoid physical scraping between heterogeneous equipment has become a pressing technical problem to be solved in this field. Summary of the Invention

[0004] This application addresses the technical problems of existing scheduling systems being detached from production business logic, failing to consider changes in the physical attributes of heterogeneous trolleys leading to delays in critical order logistics, and frequent deadlocks in complex channels. It provides a method, device, and medium for dynamic path planning of autonomous mobile robots that combines production business time windows with dynamic physical envelope constraints.

[0005] To address the aforementioned technical problems, this application provides a dynamic path planning method for autonomous mobile robots, including...

[0006] A planning map based on a three-dimensional spatiotemporal grid matrix is ​​established, wherein the three-dimensional spatiotemporal grid matrix includes a first physical coordinate dimension X, a second physical coordinate dimension Y, and a discrete-time slice dimension T; manufacturing execution business data is received, and a unique identifier for the target trolley is obtained through an identification module. Based on the unique identifier, the physical three-dimensional geometric dimensions and load parameters of the target trolley are extracted; a comprehensive cost evaluation calculation is performed on candidate nodes n in the planning map to obtain the evaluation result. The comprehensive cost function used in the comprehensive cost evaluation calculation includes the baseline time cost, heuristic estimated time cost, business priority penalty coefficient, and mechanical envelope interference penalty; wherein the calculation... The calculation of the business priority penalty coefficient includes determining the depletion time threshold based on the remaining buffer space of the endpoint workstation and the material consumption rate, and assigning a preset step penalty value to the business priority penalty coefficient when the estimated arrival time of the candidate node n exceeds the depletion time threshold; the calculation of the mechanical envelope interference penalty includes determining the fan-shaped spatial grid set occupied by the target trolley in the discrete time slice dimension T based on the physical three-dimensional geometric dimensions and the feedforward rotation angle, and setting the mechanical envelope interference penalty to an infinite value when there is a locked grid in the fan-shaped spatial grid set; and generating a conflict-free scheduling path sequence based on the evaluation results and sending it to the handling robot controller.

[0007] Furthermore, determining the depletion time threshold includes...

[0008] Obtain the remaining amount of the endpoint buffer recorded in the manufacturing execution business data at the current moment; obtain the material consumption rate of the corresponding process; obtain the basic running time by dividing the remaining amount of the endpoint buffer by the material consumption rate; determine the depletion time threshold based on the sum of the current moment and the basic running time.

[0009] Furthermore, the calculation of the penalty coefficient for this business priority also includes

[0010] Calculate the time difference between the estimated arrival time of the candidate node n and the depletion time threshold; determine whether the time difference is greater than zero; if the time difference is greater than zero, trigger a business delay alarm and add the step penalty value to the comprehensive cost function; if the time difference is less than or equal to zero, configure the corresponding basic priority coefficient according to the work order priority level contained in the manufacturing execution business data, and use the basic priority coefficient as the business priority penalty coefficient.

[0011] Furthermore, the set of sector-shaped spatial grids occupied by the target trolley within the discrete-time slice dimension T is determined to include...

[0012] Based on the trolley length, trolley width, and feedforward rotation angle in the physical three-dimensional geometry, an envelope polygon is constructed on a two-dimensional plane consisting of the first physical coordinate dimension X and the second physical coordinate dimension Y. The two-dimensional grid covered by the envelope polygon on the two-dimensional plane is expanded outward on one side to obtain an expanded grid set. The expanded grid set is combined with the discrete time slice dimension T to generate a fan-shaped spatial grid set in the three-dimensional spatiotemporal dimension.

[0013] Furthermore, the two-dimensional mesh covered by the envelope polygon on the two-dimensional plane is expanded outward on one side, including...

[0014] Read the current running speed and load parameters of the target trolley; look up the corresponding inertial offset distance compensation amount in the table according to the product of the running speed and load parameters; add the preset basic expansion radius and the inertial offset distance compensation amount to obtain the dynamic expansion radius; take the geometric center of the envelope polygon as the reference, extend the normal of each edge line of the envelope polygon according to the dynamic expansion radius, and determine the extended expansion mesh set.

[0015] Furthermore, the unique identifier of the target trolley is obtained through the identification module, including...

[0016] The radio frequency (RF) query signal is transmitted using an RF reader mounted on the transport robot body; the RF response signal containing the unique identifier is received from the passive electronic tag attached to the target trolley; the RF response signal is analyzed and the unique identifier is extracted.

[0017] Furthermore, after generating a conflict-free scheduling path sequence based on the evaluation results and sending it to the handling robot controller, it also includes...

[0018] The three-dimensional spatiotemporal occupancy information corresponding to the conflict-free scheduling path sequence is broadcast to the entire network in the pre-constructed spatiotemporal planning tree; the grid status corresponding to the three-dimensional spatiotemporal occupancy information is updated to the locked state; path query requests from other handling robots are received, and path query requests that cause interference are rejected based on the locked state.

[0019] Furthermore, the refusal to generate interference from path query requests based on the locked state also includes...

[0020] Identify the low-priority parameters of the blocked robot that initiated the path query request; calculate the waiting time cost of the blocked robot based on the locked state; when the waiting time cost exceeds the preset yield threshold, issue a forced yield command to the blocked robot; drive the blocked robot to perform backtracking and replanning based on the planning map containing the locked state.

[0021] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the dynamic path planning method described above.

[0022] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the dynamic path planning method described above.

[0023] By bidirectionally coupling underlying business data and physical parameters, this application realizes the transformation from distance-based planning to production cycle planning, and eliminates the phenomenon of high-priority order materials being blocked by low-priority tasks based on time window prediction; through a three-dimensional spatiotemporal grid locking mechanism combined with the calculation of the trolley's exclusive dynamic expansion envelope surface, it effectively intercepts path queries that generate spatial overlap, and completely eliminates the physical scraping risk and deadlock hidden danger of heterogeneous long-axis trolleys when turning or intersecting. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the system architecture provided in an embodiment of the present invention.

[0025] Figure 2 This is a flowchart of the method provided in an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram illustrating the calculation principle of the dynamic expansion envelope surface provided in an embodiment of the present invention.

[0027] Figure 4 This is a schematic diagram of the overall structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the specific embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0029] Reference Figure 1 , Figure 1 This is a schematic diagram of the system architecture provided in an embodiment of this application. This embodiment provides an autonomous mobile robot dynamic path planning system based on business timing and trolley attributes, deployed in a flexible logistics scenario within discrete manufacturing. The system architecture mainly includes a central scheduling server 101, a manufacturing execution data interface node 102, an autonomous mobile robot 103, and a target trolley 104.

[0030] In the physical infrastructure, a dot-matrix two-dimensional location tag network is deployed on the factory ground to provide absolute coordinate references for the autonomous mobile robot 103. Ultra-high frequency passive tags 106 are rigidly fixed to the bottom structure of the target trolley 104. The central scheduling server 101 houses a database module containing a business state machine and a spatiotemporal path engine module. The database module synchronizes in real-time the work order scheduling sequence pushed by the manufacturing execution data interface node 102 and the material consumption rate of each workstation. When a scheduling task is triggered, the central scheduling server 101 does not directly perform a shortest path search based on geometric distance, but first establishes a mapping relationship between the order flow and the logistics entity. The autonomous mobile robot 103 integrates a low-level motion controller and a wireless communication module inside its body. This wireless communication module can establish a wireless local area network or 5G bidirectional data transmission with the central scheduling server 101. An RFID reader 105, such as an RFID read / write antenna, is integrated on the upper surface of the autonomous mobile robot 103's body. When the autonomous mobile robot 103 moves under the target trolley 104 to perform lifting or traction actions, the RFID reader 105 can read the UHF passive tag 106, thereby initiating dynamic spatiotemporal path planning. Through this data flow structure of server coordination and edge device feedback, real-time integration of upper-layer business instructions and lower-layer physical characteristics can be ensured.

[0031] Reference Figure 2 , Figure 2 This is a flowchart of the method provided in this application embodiment. The main process of this embodiment is executed by the aforementioned central scheduling server 101 and its internal spatiotemporal path engine module, and mainly includes the following steps:

[0032] In step S201, a planning map based on a three-dimensional spatiotemporal grid matrix is ​​established. This three-dimensional spatiotemporal grid matrix includes a first physical coordinate dimension X, a second physical coordinate dimension Y, and a discrete time slice dimension T. The spatiotemporal path engine module maps the physical factory area into a tensor space with a time axis. The parameter configuration of this matrix is ​​precisely defined: the first physical coordinate dimension X and the second physical coordinate dimension Y represent two-dimensional planar physical coordinates, with a resolution set to 0.25 meters; the discrete time slice dimension T represents a future time prediction slice, with a resolution set to 0.5 seconds. Through the set high-resolution discretization processing, the robot's continuous motion trajectory can be divided into a series of finite three-dimensional state units, thereby providing a standardized data base for subsequent conflict detection.

[0033] In step S202, manufacturing execution business data is received, and the unique identifier of the target trolley 104 is obtained through the identification module. Based on the unique identifier, the physical three-dimensional geometric dimensions and load parameters of the target trolley 104 are extracted. The RFID reader 105 on the autonomous mobile robot 103 transmits an RFID query signal and receives the RFID response signal containing the unique identifier from the UHF passive tag 106 attached to the target trolley 104. The RFID response signal is parsed and the unique identifier is extracted. Based on the unique identifier, the central scheduling server 101 extracts the three-dimensional geometric dimensions (including trolley length L, width, and center of mass height) and load level of the physical trolley from the preset trolley digital twin parameter table. The extracted data is transmitted back via a wireless local area network and input into the spatiotemporal path engine module. By using the configuration of near-field extraction of trolley characteristics using RFID hardware, discrete heterogeneous trolley physical data can be accurately introduced into the scheduling backend, avoiding the computational blind spots caused by treating all vehicles as uniform-sized points in traditional control.

[0034] In step S203, a comprehensive cost evaluation calculation is performed on the candidate node n in the planning map to obtain the evaluation result. The comprehensive cost function used in the comprehensive cost evaluation calculation includes the baseline time cost, the heuristic estimated time cost, the business priority penalty coefficient, and the mechanical envelope interference penalty. The engine performs spatial search based on a three-dimensional spatiotemporal grid matrix of dimension (X,Y,T). In order to find a path that balances time optimization and business optimization, this embodiment constructs a composite cost function. The system performs an operation to add the baseline time cost, the heuristic estimated time cost, the business priority penalty coefficient, and the mechanical envelope interference penalty to obtain the comprehensive evaluation cost. The calculation logic is implemented through the following formula:

[0035] .

[0036] Wherein, variable F(n) represents the comprehensive evaluation cost of the current candidate node n; variable n represents any discrete state node to be explored in the 3D spatiotemporal map; variable G(n) represents the baseline time cost consumed from the starting node to the current candidate node n, i.e., the actual movement path time; variable H(n) represents the heuristic estimated time cost from the current candidate node n to the target node, usually configured as the estimated time value corresponding to the Manhattan distance from the candidate node n to the target node; variable P biz (n) represents the business priority penalty coefficient; variable P mech (n) represents the mechanical envelope interference penalty.

[0037] For the business priority penalty coefficient P bizThe calculation of (n) is based on the core logic of driving changes in time cost by analyzing the material shortage at the final workstation. The central scheduling server 101 obtains the remaining amount M of the final workstation buffer recorded in the manufacturing execution business data at the current moment. R Simultaneously, obtain the material consumption rate V of the corresponding process. C The spatiotemporal path engine performs division operations based on workstation data. The system divides the remaining amount in the buffer area of ​​the endpoint workstation by the material consumption rate to obtain the depletion time threshold. The calculation logic is implemented through the following formula:

[0038]

[0039] Among them, variable T D The depletion time threshold represents the maximum remaining time that the existing materials at the target workstation can support the continuous operation of the production line; variable M R Represents the remaining quantity; variable V C This indicates the number of items consumed per unit of time. Based on the sum of the current absolute system time and the calculated base running time (i.e., the running period required for the reserve to be exhausted), the system establishes an absolute depletion time threshold timestamp.

[0040] Subsequently, the system calculates the time difference between the estimated arrival time of candidate node n and the depletion time threshold, and determines whether the time difference is greater than zero. If the time difference is greater than zero, a service delay alarm is triggered, and P is assigned to the comprehensive cost function. biz (n) A step-maximum penalty value forces the algorithm to abandon the current path and search for a shorter path branch or forces other vehicles to give way. If the time difference is less than or equal to zero, it means that the current delivery time is within the safety window. The system configures the corresponding basic priority coefficient according to the work order priority level contained in the business data and uses it as the normal P. biz (n) value. Through this business penalty logic, higher passage privileges can be granted to the high material shortage state based on the actual production rhythm, realizing a paradigm upgrade from traditional spatial scheduling to business time-series scheduling.

[0041] For mechanical envelope interference penalty P mech The core logic of calculating (n) is to expand and project the physical dimensions of the trolley within a spatiotemporal slice. (Refer to...) Figure 3 , Figure 3This is a schematic diagram illustrating the calculation principle of the dynamic expansion envelope surface provided in this application embodiment, with a coordinate system established at the origin O. Based on the trolley length L, trolley width, and feedforward rotation angle θ in the physical three-dimensional geometry, the system constructs a basic two-dimensional boundary 301 on a two-dimensional plane. Due to the influence of inertia and centripetal force on the trolley during movement, the system further reads the current running speed and load parameters of the target trolley, and obtains the corresponding inertial offset distance compensation amount by looking up the product of these two parameters. The preset basic expansion radius is added to the inertial offset distance compensation amount to obtain the dynamic expansion radius. Using the geometric center of the basic two-dimensional boundary 301 as a reference, each edge line is extended along its normal direction according to the dynamic expansion radius, thereby obtaining the outwardly expanded expansion boundary 302. This constitutes an expansion mesh set on the two-dimensional plane. This expansion mesh set is projected and accumulated within the discrete-time slice dimension T to generate a fan-shaped spatial mesh set Set (Grid) in the three-dimensional spatiotemporal dimension. When the spatiotemporal path engine traverses candidate nodes, if it finds a grid state within the fan-shaped spatial grid set Set(Grid) that has been locked by other vehicles, it determines that physical interference has occurred. At this time, the system will apply a mechanical envelope interference penalty P. mech (n) is set to infinity. For scenarios where heavy-duty long-axis trolleys carrying urgent orders enter restricted passageways, the width of the grid area locked by this expanded boundary 302 can reach 1.5 times the width of a standard autonomous mobile robot. By employing a dedicated envelope surface to calculate details, the kinematic swing or braking safety distance can be pre-mapped into the map matrix, completely eliminating the risk of edge sweeping and scraping caused by heterogeneous trolleys during complex turns or intersections.

[0042] In step S204, a conflict-free scheduling path sequence is generated based on the evaluation results and sent to the transport robot controller. The central scheduling server 101 finds the sequence that minimizes the sum of F(n) in the 3D spatial planning tree, broadcasts its corresponding 3D spatiotemporal occupancy information across the entire network, and updates the (X,Y,T) grid states along the route to a locked state. Subsequently, the underlying motion controller receives the path sequence and outputs the corresponding drive electrical signals to the wheel drive motors to drive the vehicle forward.

[0043] When space resource contention occurs in the network, the system initiates a forced avoidance and replanning mechanism. The server receives path query requests from other autonomous mobile robots. If a query request covers a grid that has already been updated to a locked state by a high-priority vehicle, the server identifies the low-priority parameters of the blocked robot that initiated the query request and calculates the waiting time cost of the blocked robot based on the locked state. When the waiting time cost exceeds a preset yielding threshold, the system immediately issues a forced yielding command to the blocked robot. In actual factory parameter configurations, for situations where large long-axle trolleys occupy aisles, the system can force low-priority vehicles carrying small empty trolleys to enter the side branch aisle 5 seconds in advance to perform retreat and replanning. Through this logic of forced broadcasting of high-priority equipment physical characteristics and low-priority equipment retreat, deadlock is prevented, giving the restricted logistics channel a stronger flexible throughput capacity. It should be noted that the above values ​​of the 5-second time threshold and the 1.5-fold expansion factor are only examples. Those skilled in the art can also achieve the technical objectives of this application by adaptively scaling based on the aisle width ratio of the actual factory site.

[0044] Reference Figure 4 , Figure 4 This is a schematic diagram of the overall structure of the electronic device provided in this application embodiment. The electronic device 401 (which may be represented as the aforementioned central dispatch server 101 or edge computing node) includes a processor 402, a memory 403, and a communication module 404. The processor 402, memory 403, and communication module 404 are electrically connected to each other via an internal bus to achieve data interaction. The memory 403 stores computer program instructions that can run on the processor 402. When the processor 402 executes the computer program, it is configured to implement all method steps, such as the comprehensive cost calculation algorithm and grid locking logic described in detail in steps S201 to S204. Those skilled in the art will understand that the communication module 404 is used to establish a radio frequency or wireless signal connection with an external vehicle-side network to issue instructions. Through the combination of hardware architectures, robust underlying computing power support can be provided for large-scale, high-concurrency 3D tensor map computation.

[0045] This application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a corresponding processor, implements all the steps of the aforementioned dynamic path planning method based on business timing and trolley attributes. The storage medium can be a non-volatile data storage medium such as a read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0046] In summary, this application's embodiments construct a time-dimensional business-level penalty by extracting the actual order material consumption rate and remaining capacity, while simultaneously utilizing radio frequency identification to obtain the heterogeneous three-dimensional dimensions of vehicles and their traction trolleys to construct a spatiotemporal-dimensional dynamic expansion penalty. This allows the multi-dimensional cost function to effectively drive high-performance and heavily loaded vehicles to pass first. This combination of features not only solves the pain point of order logistics delays in traditional technologies but also addresses the difficulty of unifying the physical dimensions of non-standard trolleys, ultimately achieving efficient system-level flow and zero-scratching safety.

[0047] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A dynamic path planning method for an AMR (Automatic Transporter) robot, characterized in that, include Establish a planning map based on a three-dimensional spatiotemporal grid matrix, wherein the three-dimensional spatiotemporal grid matrix includes a first physical coordinate dimension X, a second physical coordinate dimension Y, and a discrete time slice dimension T; Receive manufacturing execution business data, obtain the unique identifier of the target trolley through the identification module, and extract the physical three-dimensional geometric dimensions and load parameters of the target trolley based on the unique identifier; A comprehensive cost assessment is performed on candidate node n in the planning map to obtain the assessment result. The comprehensive cost assessment calculation uses a comprehensive cost function that includes baseline time cost, heuristic estimated time cost, business priority penalty coefficient, and mechanical envelope interference penalty. The calculation of the business priority penalty coefficient includes determining the depletion time threshold based on the remaining amount of the end station buffer and the material consumption rate, and assigning a preset step penalty value to the business priority penalty coefficient when the estimated time to reach the candidate node n exceeds the depletion time threshold. The calculation of the mechanical envelope interference penalty includes determining the fan-shaped spatial grid set occupied by the target trolley in the discrete time slice dimension T based on the physical three-dimensional geometric dimensions and the feedforward rotation angle, and setting the mechanical envelope interference penalty to an infinite value when there is a locked grid in the fan-shaped spatial grid set; Based on the evaluation results, a conflict-free scheduling path sequence is generated and sent to the transport robot controller.

2. The method as described in claim 1, characterized in that, Determining the depletion time threshold includes Get the remaining amount of the endpoint workstation buffer recorded in the manufacturing execution business data at the current moment; Obtain the material consumption rate for the corresponding process; The basic running time is obtained by dividing the remaining amount in the buffer area of ​​the final workstation by the material consumption rate. The depletion time threshold is determined based on the sum of the current time and the base running time.

3. The method as described in claim 2, characterized in that, Calculating the penalty coefficient for this business priority also includes Calculate the time difference between the estimated arrival time of the candidate node n and the depletion time threshold; Determine if the time difference is greater than zero; If the time difference is greater than zero, a business delay alarm will be triggered, and the step penalty value will be added to the comprehensive cost function. If the time difference is less than or equal to zero, the corresponding basic priority coefficient is configured according to the work order priority level contained in the manufacturing execution business data, and the basic priority coefficient is used as the penalty coefficient for the priority of that business.

4. The method as described in claim 1, characterized in that, The set of sector-shaped spatial grids occupied by the target trolley within the discrete-time slice dimension T includes... Based on the trolley length, trolley width, and feedforward rotation angle in the physical three-dimensional geometry, an envelope polygon is constructed on a two-dimensional plane consisting of the first physical coordinate dimension X and the second physical coordinate dimension Y. The two-dimensional mesh covered by the envelope polygon on the two-dimensional plane is expanded outward on one side to obtain an expanded mesh set; By combining the dilated mesh set with the discrete-time slice dimension T, a fan-shaped spatial mesh set in the three-dimensional spatiotemporal dimension is generated.

5. The method as described in claim 4, characterized in that, Expanding the 2D mesh covered by the envelope polygon on the 2D plane outwards on one side, including... Read the current operating speed and load parameters of the target trolley; The corresponding inertial offset distance compensation amount is obtained by looking up a table based on the product of the operating speed and load parameters; The dynamic expansion radius is obtained by adding the preset basic expansion radius to the inertial offset distance compensation amount; Using the geometric center of the envelope polygon as a reference, extend the normal of each edge line of the envelope polygon according to the dynamic expansion radius to determine the extended expansion mesh set.

6. The method as described in claim 1, characterized in that, The unique identifier of the target trolley is obtained through the identification module, including... Radio frequency query signals are transmitted using an RFID reader mounted on the body of the transport robot; Receive the radio frequency response signal containing the unique identifier fed back by the passive electronic tag attached to the target trolley; The radio frequency response signal is analyzed and the unique identifier is extracted.

7. The method as described in claim 1, characterized in that, After generating a conflict-free scheduling path sequence based on the evaluation results and sending it to the handling robot controller, it also includes... The three-dimensional spatiotemporal occupancy information corresponding to the conflict-free scheduling path sequence is broadcast across the entire network in a pre-constructed spatiotemporal planning tree; Update the grid state corresponding to the 3D spatiotemporal occupancy information to the locked state; It receives path query requests from other handling robots and rejects path query requests that would cause interference based on the locked state.

8. The method as described in claim 7, characterized in that, Refusing to interfere with path query requests based on a locked state also includes... Identify the low-priority parameters of the blocked robot that initiated the path query request; Calculate the waiting time cost of the blocked robot based on the locked state; When the waiting time cost exceeds the preset yield threshold, a forced yield command is issued to the blocked robot. The blocked robot is driven to perform retreat and replanning based on a planning map that includes the locked states.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in claim 1.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When a computer program is executed by a processor, it implements the method as described in claim 1.

Images