A multi-floor building heterogeneous robot scheduling method and system and a computer readable storage medium

Abstract

This invention discloses a method, system, and computer-readable storage medium for scheduling heterogeneous robots in multi-story buildings. The method acquires a 3D road network including planar access edges, unidirectional vertical edges of staircases, and virtual composite resources of elevators. It performs candidate path pre-calculation for a set of heterogeneous robots and tasks, generating a set of the top K candidate paths for robot-task pairings, and obtains the basic time consumption for pairing based on the optimal path. A bipartite graph is constructed using the basic time consumption as edge weights, and the Kuhn-Munkres algorithm is used to complete the matching solution, deriving the initial spatiotemporal occupancy scheme. Multi-dimensional conflict detection is performed on the initial scheme, considering edge and node capacity, elevator lobby and car capacity, and delivery / pickup timing constraints, to obtain a conflict set. The final scheduling scheme is then output. This invention pre-calculates paths and uses path replacement rather than simple delay waiting to resolve conflicts, which can identify vertical resource conflicts in multi-story buildings, reduce online scheduling computational overhead, and decrease task latency.

Description

Technical Field

[0001] This invention relates to the field of robot scheduling technology, and in particular to a method, system, and computer-readable storage medium for scheduling heterogeneous robots in multi-story buildings. Background Technology

[0002] As heterogeneous robot swarms are gradually deployed in multi-story indoor environments such as hospitals and large buildings, utilizing multiple heterogeneous robots with varying motion performance and mobility to complete material delivery can effectively improve the efficiency of material transfer within buildings. Multi-story building environments contain various access resources, including horizontal corridors, staircases, elevator lobbies, and elevator cars. Staircases and elevators are particularly scarce vertical access resources. Robots are highly susceptible to resource contention, queuing congestion, and even deadlocks during cross-floor operations, placing high demands on the entire process of robot swarm scheduling, including path pre-planning, task allocation, and conflict resolution.

[0003] Most existing multi-robot scheduling solutions are designed for single-level warehouse scenarios. In the path planning stage, traditional multi-robot scheduling primarily uses A* and Dijkstra's algorithms to solve for a single optimal path, lacking a mechanism for pre-generating a set of alternative paths in batches. Even when some studies employ the K-shortest path algorithm, alternative paths are often calculated online in real-time after a conflict occurs, concentrating computational overhead during the scheduling operation phase. When the robot cluster is large, online real-time pathfinding places enormous computational pressure on the system, impacting its response performance.

[0004] In terms of task allocation, existing scheduling systems often separate task allocation from path planning. Some robot task matching and scheduling schemes based on bipartite graphs only match robots and tasks based on path travel time, lacking data linkage with the pre-calculated set of multiple alternative paths. After task matching, a single optimal path is usually used directly for time-series extrapolation, which can easily lead to a large number of robots simultaneously vying for scarce vertical resources such as elevators and stairs. This results in a large number of potential conflicts in the generated initial scheduling scheme, placing a significant burden on subsequent conflict resolution processes.

[0005] At the conflict resolution level, mainstream multi-robot conflict resolution methods mainly rely on delayed waiting and deceleration avoidance, only performing time window conflict detection on planar road networks and only handling time sequence offsets for conflicting robots. This approach does not change the robot's travel path. When a large number of robots compete for limited vertical resources, simply relying on waiting and avoidance will significantly increase task completion time and may even lead to cluster deadlock. A few scheduling schemes with path replacement capabilities cannot fully consider the combined resource capacity constraints of planar passage edges, stair vertical edges, and elevators. They also lack comprehensive verification of pickup and delivery timing, load, and task time windows, resulting in incomplete conflict identification dimensions. After completing path replacement, they do not synchronously update a full set of related variables such as elevator occupancy and pickup and delivery timing, making the output scheduling scheme prone to temporal logic contradictions and poor practical executability.

[0006] In summary, existing technologies lack a complete scheduling pipeline, making it impossible to achieve effective linkage of batch pre-generation of alternative paths, global matching of robots and tasks, multi-dimensional conflict identification, and greedy conflict resolution based on alternative path sets. This makes it difficult to meet the actual needs of efficient and stable operation of heterogeneous robot clusters in multi-story building scenarios such as hospitals.

[0007] The disclosure of the above background technical content is only for the purpose of assisting in understanding the concept and technical solution of this application, and does not necessarily provide technical instruction. Summary of the Invention

[0008] The purpose of this invention is to provide a method, system, and computer-readable storage medium for scheduling heterogeneous robots in multi-story buildings.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for scheduling heterogeneous robots in multi-story buildings includes the following steps: Obtain a three-dimensional road network for a multi-story building. The three-dimensional road network includes planar passage edges, staircase unidirectional vertical edges, and elevator virtual composite resources. The elevator virtual composite resources include elevator lobby nodes and elevator car nodes. For each heterogeneous robot in the heterogeneous robot set and each task in the task set, a candidate path pre-calculation step is performed: based on the Yen algorithm, combined with the edge time corresponding to each heterogeneous robot, a set of the top K candidate paths for pairing each heterogeneous robot and task is generated, where K is a positive integer greater than or equal to 2; the optimal path in the set of the top K candidate paths is selected, and the edge time of each edge on the optimal path is accumulated to obtain the base time for pairing each heterogeneous robot and task; A bipartite graph is constructed, with the heterogeneous robot as the first type of node and the task as the second type of node. The basic time consumption for pairing the heterogeneous robot and the task is used as the edge weight of the bipartite graph. The Kuhn-Munkres algorithm is used to solve the matching problem to obtain the matching result of the heterogeneous robot and the task. Based on the matching result, the corresponding optimal path is selected, and the initial spatiotemporal occupancy scheme is generated by deducing according to discrete time steps. Multi-dimensional conflict detection is performed on the initial spatiotemporal occupancy scheme. The multi-dimensional conflict detection includes edge and node capacity conflict detection, elevator lobby and car capacity conflict detection, and delivery and pick-up time sequence constraint conflict detection to obtain a conflict set. The conflict resolution process is executed repeatedly until the conflict set is empty. The conflict resolution process includes: when the conflict set is not empty, for the conflict robot in the conflict set, the set of the first K candidate paths corresponding to the conflict robot is reused, the candidate paths are traversed, the candidate path that minimizes the global completion time increment is selected to replace the original path, the variables corresponding to the time and space occupancy scheme are updated synchronously, and the multi-dimensional conflict detection is re-executed to obtain the updated conflict set. Output the final scheduling scheme after eliminating all conflicts.

[0010] Furthermore, based on any one or a combination of the aforementioned technical solutions, in generating the preceding... K After the candidate path set is completed, a path validity filtering step is also included: Each path output by the Yen algorithm is verified, and paths containing vertical resources that are inaccessible to the heterogeneous robots are removed based on the access permissions of each heterogeneous robot.

[0011] Furthermore, following any one or a combination of the aforementioned technical solutions, the path validity filtering step further includes: calculating the similarity between paths and removing redundant similar paths whose similarity exceeds a preset threshold. Furthermore, when constructing the edge weights of the bipartite graph using any one or a combination of the aforementioned technical solutions, at least one correction cost is also superimposed. The correction cost includes task priority correction cost, heterogeneous robot adaptation cost, and vertical resource occupation penalty cost.

[0012] Furthermore, after obtaining the conflict set by any one or a combination of the aforementioned technical solutions, the process further includes a conflict classification step: Each conflict within the conflict set is classified into a danger level, and conflicts of different levels are processed sequentially according to a preset resolution priority order.

[0013] Furthermore, following any one or a combination of the aforementioned technical solutions, when all the preceding technical solutions corresponding to the conflict robot are...K If conflicts still exist after the candidate path replacement is completed, a fallback operation is performed, which includes at least one of time offset waiting, task reallocation, and task deferred queuing.

[0014] Furthermore, following any one or a combination of the aforementioned technical solutions, after replacing the alternative path and updating the elevator-related occupancy variables, an elevator timing consistency check is performed. If a timing logic contradiction is detected between the elevator lobby and the car, the corresponding waiting time or door opening / closing occupancy time is corrected.

[0015] Furthermore, following any one or a combination of the aforementioned technical solutions, the conflict resolution process is also configured with a maximum iteration threshold; if the number of iterations reaches the maximum iteration threshold and all conflicts are still not eliminated, the loop is terminated.

[0016] According to another aspect of the present invention, a multi-story building heterogeneous robot scheduling system is provided for implementing the multi-story building heterogeneous robot scheduling method as described in any of the preceding claims, comprising: The road network acquisition module is used to acquire the three-dimensional road network of multi-story buildings; The candidate path generation module is used to perform the candidate path pre-calculation step and output the set of the top K candidate paths for robot and task pairing, as well as the basic time consumption. The task matching module is used to construct a bipartite graph, call the Kuhn-Munkres algorithm to obtain the matching results between the robot and the task, and generate an initial spatiotemporal occupancy scheme. The conflict detection and resolution module is used to perform multi-dimensional conflict detection, cyclically execute the conflict repair process, and output the final scheduling scheme.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided thereon storing a computer program that, when executed by a processor, implements the multi-story building heterogeneous robot scheduling method as described in any of the preceding claims.

[0018] The beneficial effects of the technical solution provided by this invention are as follows: a. By pre-compiling multiple candidate paths for robot-task pairing, the computational overhead of path search is shifted to the preprocessing stage before scheduling. In subsequent task matching and conflict resolution stages, the pre-computed candidate paths can be directly reused, eliminating the need to call the pathfinding algorithm in real-time during scheduling. Compared to solutions that only find alternative paths online after a conflict occurs, this significantly reduces the real-time computational load during the scheduling cycle and improves the response performance of the scheduling system in large-scale heterogeneous robot cluster scenarios. b. Using the pre-calculated baseline time for robot-task pairing as input for bipartite graph matching, a global one-to-one optimal match between the robot and the task is achieved. Temporal deduction is performed based on a pre-generated set of candidate paths to avoid the disconnect between task allocation and path selection capabilities, alleviate the phenomenon of a large number of robots vying for scarce vertical resources such as stairs and elevators, reduce the number of conflicts in the initial solution, and lower the processing burden of subsequent conflict resolution stages. c. It considers multi-dimensional conflict detection, taking into account edge-node capacity, elevator lobby and car resources, delivery and pickup timing, load, and task time windows. When a conflict occurs, it prioritizes reusing the pre-computed candidate path set, selecting the path with the smallest global completion time increment to complete the replacement, and simultaneously updating the elevator occupancy and the entire set of delivery and pickup timing variables. Unlike traditional conflict handling methods that simply delay and wait, this reduces task latency caused by mere waiting, lowers the risk of cluster congestion and deadlock, and ensures stable and efficient operation of heterogeneous robot clusters in multi-story buildings. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram of a multi-story building heterogeneous robot scheduling method is provided as an exemplary embodiment of the present invention; Figure 2 This is a schematic diagram of a multi-story building heterogeneous robot scheduling system, provided as an exemplary embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0023] In one embodiment of the present invention, a method for scheduling heterogeneous robots in multi-story buildings is provided, see [link to relevant documentation]. Figure 1 The process includes the following steps: obtaining a three-dimensional road network of a multi-story building, wherein the three-dimensional road network includes planar passage edges, staircase unidirectional vertical edges, and elevator virtual composite resources, wherein the elevator virtual composite resources include elevator lobby nodes and elevator car nodes; For each heterogeneous robot in the set of heterogeneous robots and each task in the set of tasks, a candidate path pre-calculation step is performed: Based on the Yen algorithm, and combined with the edge time corresponding to each heterogeneous robot, a set of the top K candidate paths for pairing each heterogeneous robot and task is generated, where K is a positive integer greater than or equal to 2; the optimal path in the set of the top K candidate paths is selected, and the edge time of each edge on the optimal path is accumulated to obtain the base time for pairing each heterogeneous robot and task; a bipartite graph is constructed, with the heterogeneous robot as the first type of node in the bipartite graph, the task as the second type of node in the bipartite graph, and the base time for pairing the heterogeneous robot and task as the edge weight of the bipartite graph, and the Kuhn-Munkres algorithm is used for matching to obtain the matching result of the heterogeneous robot and task; based on the matching... The optimal path is selected based on the allocation result, and an initial spatiotemporal occupancy scheme is generated by deducing the discrete time step. Multi-dimensional conflict detection is then performed on the initial spatiotemporal occupancy scheme, including edge and node capacity conflict detection, elevator lobby and car capacity conflict detection, and delivery / pickup timing constraint conflict detection, resulting in a conflict set. A conflict repair process is executed iteratively until the conflict set is empty. The conflict repair process includes: when the conflict set is not empty, for the conflicting robots within the conflict set, the set of the top K candidate paths corresponding to the conflicting robots is reused, the candidate paths are traversed, and the candidate path that minimizes the global completion time increment is selected to replace the original path. The variables corresponding to the spatiotemporal occupancy scheme are updated synchronously, and the multi-dimensional conflict detection is re-executed to obtain the updated conflict set. Finally, the final scheduling scheme after eliminating all conflicts is output.

[0024] This embodiment pre-generates a set of the top K candidate paths for heterogeneous robot and task pairing, providing a library of alternative paths for subsequent conflict resolution. This pre-processing of path generation eliminates the need for real-time pathfinding calculations during scheduling. Based on the pre-calculated baseline time, bipartite graph matching is performed to complete the global pairing of heterogeneous robots and tasks and deduce the initial spatiotemporal occupancy scheme. Conflict detection is conducted from multiple dimensions, including edge and node capacity, elevator lobby and car capacity, and delivery / pickup sequence constraints, identifying various conflicts arising from planar and vertical access resources in multi-story building scenarios. When a conflict occurs, the pre-generated candidate path set is reused, and the alternative path with the smallest global completion time increment is selected to replace the original path. Conflict detection and resolution are iteratively executed until the conflict is eliminated, ultimately outputting a conflict-free scheduling scheme. Compared to existing technologies where task allocation and conflict resolution are separated, conflict handling relies solely on delays, and vertical resource conflict identification is lacking, this approach reduces computational overhead during online scheduling, alleviates vertical resource contention, and lowers task latency caused by simple waiting.

[0025] In one embodiment of the present invention, a method for scheduling heterogeneous robots in multi-story buildings is provided, see [link to relevant documentation]. Figure 1 This includes the following steps: A 3D road network of a multi-story building is obtained. This network includes planar travel edges, unidirectional vertical stair edges, and virtual composite elevator resources, which include elevator lobby nodes and elevator car nodes. The 3D road network outputs the edge latency for each edge corresponding to different heterogeneous robots. , k Let 'e' be the robot's ID number, and 'e' be an edge in the road network. Indicates the first k The inherent travel time of a robot on edge e does not include the additional time caused by resource waiting or congestion.

[0026] Perform candidate path pre-computation: For each robot in the robot set and each task in the task set, generate the first candidate paths for robot-task pairings. K A set of candidate paths. For the... k Taiwan Robot With the j-th task : For the task Pick-up points, For the task Delivery nodes; based on the Yen algorithm, with As edge weights, the solution from the pickup node in the 3D road network is obtained. To delivery node The former KEach path is analyzed to obtain a path list PathList, which is then saved as a set of candidate paths for that pair. .

[0027] The optimal path ranked first in the candidate path set PathList [0]Calculate the base time for robot and task pairing Calculation formula:

[0028] in: Represents robots Execute the task The base execution time; PathList[0] represents the previous time. K The optimal path among the candidate paths; e∈PathList[0] represents each road network edge contained in the optimal path; Represents robots The inherent travel time on edge e.

[0029] Iterate through all robots and all tasks to obtain a set of candidate paths for matching all robots and tasks. Compared with the base time .

[0030] Robot and task binary matching assignment: Construct a bipartite graph where the first type of nodes are robots in robot set M, and the second type of nodes are tasks in task set J; the base time for matching robots and tasks is used. As the weights of the corresponding edges in the bipartite graph, the Kuhn-Munkres algorithm is called to find the minimum total time required for the matching result, Match.

[0031] Initialize the allocation matrix W The matrix elements are , The value can be 0 or 1; if the robot With the task If a match is made in the Match results, then set... =1, otherwise =0; The constraint condition is satisfied:

[0032] in: To assign elements to the matrix, =1 means the task will be completed. Assigned to robots ; Indicates for any task Only one robot is assigned this task.

[0033] right When a robot with a value of 1 is paired with a task, the corresponding set of candidate paths is read. Select the optimal path within the set; perform time series analysis according to discrete time step T, and fill the edge occupancy variables. Node occupancy variables ;in Representative moment t robot k Does it occupy an edge? e , Representative moment t robot k Does it occupy a node? v At the same time, based on the task pickup time window [ Constraint Pickup Timing Mark Delivery sequence marker ; For the task The earliest allowed pickup time, For the task The latest permitted pickup time; The pickup was completed at time t. The delivery is completed at time t. After the deduction is completed, the initial spatiotemporal occupancy scheme is obtained. }

[0034] Execute a greedy spatiotemporal conflict resolution scheduling: Read the initial spatiotemporal occupancy scheme and road network edge capacity parameters. Node capacity parameters Elevator capacity parameters (Lobby capacity) (Car capacity); Continuously carry out conflict detection and conflict repair until there are no more conflicts.

[0035] The first step is to perform multi-dimensional conflict detection and initialize the conflict set. .

[0036] Perform edge and node capacity conflict detection: traverse all discrete time steps. t All edges e All nodes v ,judge:

[0037] in: For the edge e The maximum number of robots that can be accommodated; For nodes v The maximum number of robots that can be accommodated. If the above formula is satisfied, then the corresponding ( k,e,t )or( k,v,t Add to conflict set Conflict .

[0038] Perform elevator lobby and car capacity conflict detection: traversal time t Each elevator r ,judge:

[0039] in: For a moment t robot k occupying the elevator r The signage in the foyer; For a moment t robot k occupying the elevator r The markings on the car; For elevator r The maximum number of robots that the lobby can accommodate; For elevator r The maximum number of robots the car can hold. If the condition is met, the corresponding ( k,r,t Add to conflict set Conflict .

[0040] Perform conflict detection for delivery / pickup timing, load capacity, and time window constraints: Verify the following constraints: delivery time is no earlier than pickup time plus pickup operation duration; the number of tasks a single robot can simultaneously handle is no greater than 1; pickup time is no earlier than the earliest pickup time 'a' of the task. If any constraint is not satisfied, add the corresponding item to the conflict set. Conflict .

[0041] Step 2, Conflict Detection: If... Exit the loop.

[0042] The third step is greedy conflict resolution: For the conflicting robots within the conflict set, read the set of candidate paths that match the robot with the corresponding task. Iterate through all candidate paths in the set; for each candidate path replaced, re-analyze the time sequence and calculate the global completion time increment. Select Replace the existing path with the smallest alternative path; synchronously update the edge occupancy variable. Node occupancy variables Elevator occupancy marker , and delivery time stamps After the replacement is complete, return to the conflict detection step and re-detect the conflict.

[0043] Once the loop ends without conflict, calculate the global completion time. :

[0044] In the formula: The global completion time for all tasks; For the task j Delivery time; =1 indicates that the robot is at time t. k Complete the task j Delivery.

[0045] Output a conflict-free final scheduling scheme. } and overall completion time .

[0046] For example, the number of candidate paths K can be 3, 4, 5, or 10; the preset threshold for path similarity can be set to 0.8, and when the similarity between two paths is greater than 0.8, they are judged as redundant similar paths and are removed; the maximum number of iterations in the conflict resolution process can be set to 20; when classifying conflict risk levels, the classification threshold can be set according to the resource occupancy level and robot priority, for example, classifying conflicts between elevator cars and staircases as high-risk, and conflicts involving ordinary corridor edge nodes as low-risk. The multi-story building heterogeneous robot scheduling method disclosed in this embodiment achieves pipelined linkage of candidate path pre-calculation, robot and task binary matching allocation, and greedy spatiotemporal conflict resolution. It moves the computational overhead of path search to the preprocessing stage, and the task matching and conflict repair stages directly reuse the pre-generated candidate path set, avoiding frequent real-time pathfinding during scheduling and reducing the computational load of online scheduling. It uses the pre-calculated basic time to complete global binary matching, optimizes the initial scheduling scheme, alleviates the competition for scarce vertical resources such as stairs and elevators, and reduces the number of initial conflicts. Furthermore, it can identify conflicts from multiple dimensions such as edge node capacity, elevator lobby and car resources, delivery and pickup sequence, and task time window. When a conflict occurs, it adopts a greedy repair strategy of alternative path replacement, which is different from the simple delay waiting method. At the same time, it synchronously updates various related time sequence variables, reduces task waiting latency, reduces the probability of robot cluster congestion and deadlock, and is suitable for the actual working conditions of large-scale collaborative scheduling of heterogeneous robot clusters in multi-story buildings.

[0047] In one embodiment of the present invention, a multi-story building heterogeneous robot scheduling system is provided, comprising: The road network acquisition module is used to acquire the three-dimensional road network of a multi-story building. The three-dimensional road network includes planar passage edges, staircase unidirectional vertical edges, and elevator virtual composite resources. The elevator virtual composite resources include elevator lobby nodes and elevator car nodes. The module also acquires the edge consumption time τ of the heterogeneous robot corresponding to each edge.

[0048] The candidate path generation module performs candidate path pre-computation: for each robot in the robot set and each task in the task set, based on the Yen algorithm, it calculates the candidate paths... Using edge weights, generate the top K candidate paths for robot-task pairing. And according to the formula:

[0049] The base time for matching robots with tasks Where PathList[0] is the optimal path in the candidate path set; Indicates the first k Taiwan robot performs the first j The base travel time for each task; output the set of candidate paths for all robot and task pairings. Basic time consumption It is transmitted to the downstream task matching module.

[0050] The task matching module, connected to the candidate path generation module, is used to perform a binary matching and allocation between the robot and the task: receiving the basic time consumption. Construct a bipartite graph, call the Kuhn-Munkres algorithm to solve for the matching results; generate the assignment matrix. W Matrix elements Satisfy constraints:

[0051] Read the candidate path set For successfully assigned robots and task pairings, the optimal path is selected, the timing sequence is deduced according to discrete time steps, and the edge occupancy variables are filled. Node occupancy variables The task time window is combined to generate pickup and delivery time sequence markers. The initial spacetime occupancy scheme is output and passed to the conflict detection and resolution module; whereby... =1 represents a task Assigned to robots , Representative moment t robot k Does it occupy an edge? e , Representative moment t robot k Does it occupy a node? v , The pickup was completed at time t. The delivery is completed at time t.

[0052] The conflict detection and resolution module, connected to the task matching module, is used to perform greedy spatiotemporal conflict resolution scheduling: it receives the initial spatiotemporal occupancy scheme, road network and elevator capacity parameters; it iteratively executes multi-dimensional conflict detection, including edge node capacity conflict detection, elevator lobby and car capacity conflict detection, pickup and delivery timing, load, and time window constraint detection, to obtain the conflict set Conflict; when the conflict set is not empty, it reads the candidate path set corresponding to the conflicting robot. Iterate through the candidate paths and select the global completion time increment. The shortest path completes the replacement, synchronously updating all temporal variables for edges, nodes, elevators, and pickup / delivery, and then re-performs conflict detection; until the conflict set is empty; according to the formula:

[0053] Calculate the global completion time and output the final conflict-free scheduling scheme and global completion time. ;in, Let e ​​be the maximum number of robots that can be accommodated. Let v be the maximum number of robots it can accommodate. This is the maximum capacity of the elevator lobby. This refers to the maximum capacity of the elevator car. , These are markers indicating where robots occupy the elevator lobby and car, respectively. The time spent on delivery operations for the task.

[0054] In one embodiment of the present invention, a computer-readable storage medium is provided. This computer-readable storage medium is applied to a multi-story building heterogeneous robot scheduling scenario, and stores computer-executable program instructions thereon. When the computer-executable program instructions are read and executed by one or more processors, all steps of the multi-story building heterogeneous robot scheduling method described in the foregoing embodiments are implemented.

[0055] Specifically, the computer-readable storage medium is a non-volatile readable storage medium, which can be a read-only memory (ROM), random access memory (RAM), flash memory, hard disk, solid-state drive, magnetic storage disk, optical disk, or any other storage medium with program storage capability. This embodiment does not limit the specific type of storage medium. Those skilled in the art should understand that the storage medium can be independently set up locally on the scheduling server or deployed in a remote storage device. The processor can call the program instructions inside the storage medium through local reading or network access, thereby completing the complete scheduling process, including candidate path pre-calculation, robot-task binary matching and allocation, and greedy spatiotemporal conflict resolution, and outputting a conflict-free robot cluster scheduling scheme to achieve efficient collaborative delivery operations of heterogeneous robots in multi-story building environments.

[0056] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0057] The above description is only a specific embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for scheduling heterogeneous robots in multi-story buildings, characterized in that, Includes the following steps: Obtain a three-dimensional road network for a multi-story building. The three-dimensional road network includes planar passage edges, staircase unidirectional vertical edges, and elevator virtual composite resources. The elevator virtual composite resources include elevator lobby nodes and elevator car nodes. For each heterogeneous robot in the heterogeneous robot set and each task in the task set, a candidate path pre-calculation step is performed: based on the Yen algorithm, and combined with the edge time consumption corresponding to each heterogeneous robot, a candidate path pre-calculation step is generated for each heterogeneous robot and task pairing. K A set of candidate paths K It is a positive integer greater than or equal to 2; select the preceding... K The optimal path in the candidate path set is obtained by summing the edge time of each edge on the optimal path to obtain the base time for each heterogeneous robot and task pairing. A bipartite graph is constructed, with the heterogeneous robot as the first type of node and the task as the second type of node. The basic time consumption for pairing the heterogeneous robot and the task is used as the edge weight of the bipartite graph. The Kuhn-Munkres algorithm is used to solve the matching problem to obtain the matching result of the heterogeneous robot and the task. Based on the matching result, the corresponding optimal path is selected, and the initial spatiotemporal occupancy scheme is generated by deducing according to discrete time steps. Multi-dimensional conflict detection is performed on the initial spatiotemporal occupancy scheme. The multi-dimensional conflict detection includes edge and node capacity conflict detection, elevator lobby and car capacity conflict detection, and delivery and pick-up time sequence constraint conflict detection to obtain a conflict set. The conflict resolution process is executed cyclically until the conflict set is empty; the conflict resolution process includes: when the conflict set is not empty, for the conflicting robots in the conflict set, reusing the previous conflicting robot corresponding to the conflicting robot. K The candidate path set is traversed, and the candidate path that minimizes the global completion time increment is selected to replace the original path. The variables corresponding to the time and space occupancy scheme are updated synchronously, and the multi-dimensional conflict detection is re-executed to obtain the updated conflict set. Output the final scheduling scheme after eliminating all conflicts.

2. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, Before generating the above K After the candidate path set is completed, a path validity filtering step is also included: Each path output by the Yen algorithm is verified, and paths containing vertical resources that are inaccessible to the heterogeneous robots are removed based on the access permissions of each heterogeneous robot.

3. The method for scheduling heterogeneous robots in multi-story buildings according to claim 2, characterized in that, The path validity filtering step further includes: calculating the similarity between paths and removing redundant similar paths whose similarity exceeds a preset threshold.

4. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, When constructing the edge weights of the bipartite graph, at least one correction cost is also superimposed, including task priority correction cost, heterogeneous robot adaptation cost, and vertical resource occupation penalty cost.

5. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, After obtaining the conflict set, the process further includes a conflict classification step: Each conflict within the conflict set is classified into a danger level, and conflicts of different levels are processed sequentially according to a preset resolution priority order.

6. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, When all the previous ones corresponding to the conflict robot K If conflicts still exist after the candidate path replacement is completed, a fallback operation is performed, which includes at least one of time offset waiting, task reallocation, and task deferred queuing.

7. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, After replacing the alternative path and updating the elevator-related occupancy variables, an elevator timing consistency check is performed. If a timing logic contradiction is detected between the elevator lobby and the car, the corresponding waiting time or door opening / closing occupancy time is corrected.

8. The method for scheduling heterogeneous robots in multi-story buildings according to claim 1, characterized in that, The conflict resolution process is also configured with a maximum iteration threshold; if all conflicts are not eliminated after the number of iterations reaches the maximum iteration threshold, the loop is terminated.

9. A multi-story building heterogeneous robot scheduling system, characterized in that, The method for scheduling heterogeneous robots in multi-story buildings as described in any one of claims 1-8 includes: The road network acquisition module is used to acquire the three-dimensional road network of multi-story buildings; The candidate path generation module is used to perform the candidate path pre-calculation step and output the set of the top K candidate paths for robot and task pairing, as well as the basic time consumption. The task matching module is used to construct a bipartite graph, call the Kuhn-Munkres algorithm to obtain the matching results between the robot and the task, and generate an initial spatiotemporal occupancy scheme. The conflict detection and resolution module is used to perform multi-dimensional conflict detection, cyclically execute the conflict repair process, and output the final scheduling scheme.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the multi-story building heterogeneous robot scheduling method according to any one of claims 1-8.

Images