A method, modeling system, and computer-readable storage medium for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings.

By constructing a three-dimensional spatiotemporal road network with dual virtual nodes in elevators and unidirectional capacity constraints in staircases, the problems of resource contention and conflict identification and heterogeneous robot differentiation in multi-robot cluster scheduling are solved. This enables precise path planning and resource management, and improves the scheduling efficiency of robot clusters in multi-story buildings.

CN123116598APending Publication Date: 2026-11-17HANGZHOU YUANJIE ENVIRONMENTAL PROTECTION TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611336076.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-31
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies cannot effectively construct multi-story building three-dimensional spatiotemporal road networks that are adapted to the collaborative scheduling of multiple robot clusters. They cannot identify conflicts in elevator and staircase resource contention and fail to distinguish the motion characteristics of heterogeneous robots, resulting in large deviations in path prediction and incomplete resource capacity parameters.

Method used

By employing elevator dual-virtual node hierarchical modeling, staircase unidirectional capacity constraints, and heterogeneous robot differentiated time consumption calculation, a three-dimensional spatiotemporal road network with complete resource capacity parameters is constructed, accurately identifying congestion conflicts among multiple robots in multi-story buildings, and supporting large-scale cluster spatiotemporal scheduling optimization.

Benefits of technology

It enables refined modeling of elevator and staircase resources in multi-story buildings, identifies robot congestion conflicts, provides accurate path planning and resource capacity parameters, and improves the scheduling efficiency and conflict detection capabilities of multi-robot clusters.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116598A_ABST
    Figure CN123116598A_ABST
Patent Text Reader

Abstract

This invention discloses a method, modeling system, and computer-readable storage medium for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings. The method includes: planar topological modeling, constructing planar nodes, splitting bidirectional corridors into unidirectional travel edges, and calculating differentiated travel times; vertical staircase channel modeling, generating unidirectional staircase edges for both upward and downward travel, calculating travel times, and configuring fixed staircase capacity; layered virtual modeling of elevator composite resources, constructing virtual nodes for the elevator lobby and car, building cross-floor composite travel edges, and calculating robot elevator travel time; and integrating and outputting a standardized three-dimensional spatiotemporal road network, differentiated travel times, and a complete set of resource capacity parameters. This invention constructs a three-dimensional spatiotemporal road network with complete resource capacity parameters through layered modeling of elevator dual virtual nodes, unidirectional capacity constraints on staircases, and differentiated travel time calculation for heterogeneous robots. It accurately identifies congestion conflicts among multiple robots in multi-story buildings and supports large-scale cluster spatiotemporal scheduling optimization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of multi-robot cluster scheduling and multi-story building robot road network modeling technology, and in particular to a method, modeling system and computer-readable storage medium for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings. Background Technology

[0002] With the rapid deployment of indoor mobile robot technology, delivery robots are being widely deployed in multi-story buildings such as hospitals, office buildings, and multi-level warehouses to handle material transfers. Multi-story buildings also contain multiple levels of passageways, staircases, and elevators. Robots need to cross floors to complete retrieval and delivery tasks, thus requiring the construction of a passageway network adapted to multi-level three-dimensional spaces to support robot path planning and multi-robot collaborative scheduling. Currently, the mainstream multi-level robot road network modeling schemes are mainly divided into two categories: one is a single-robot autonomous mapping scheme, and the other is a simplified modeling scheme for traditional multi-robot two-dimensional road networks. Both schemes have significant technical shortcomings, as detailed below: The first type is the single-robot grid-based autonomous mapping solution: This solution relies on the robot's laser and vision sensors to collect global 3D point cloud data. After cutting the point cloud by floor, it projects it to generate a 2D grid map. Only elevator and stairwell entrances are extracted as simple connecting nodes to build the inter-floor topology. The overall solution only serves the autonomous localization and path planning of a single robot and has several inherent shortcomings: First, the basic map unit is a uniform grid, which cannot distinguish business anchor points such as wards, pharmacies, and elevator lobbies, making it difficult to build a structured topology network for material delivery business; Second, elevators and stairwells are only abstracted as a single connecting edge between two points, using a black-box fixed time parameter. First, it fails to distinguish between two independent resources: elevator lobby and elevator car, making it impossible to simulate congestion scenarios such as robot queuing and car overloading. Second, it does not differentiate between independent uphill and downhill passages on stairs, and does not set passage capacity constraints, making it impossible to identify bidirectional collisions between two robots on stairs. Third, the grid passage cost is uniformly applied globally, failing to differentiate between the planar travel speeds of different models and load-bearing heterogeneous robots, as well as the differences in low-speed travel on stairs, resulting in significant deviations in path travel time estimation. Fourth, it lacks a standardized system of edge, node, and elevator resource capacity parameters, which can only support single-machine pathfinding and cannot provide underlying quantitative data support for spatiotemporal conflict detection and global scheduling optimization for multi-robot clusters.

[0003] The second category includes traditional simplified multi-robot road network modeling schemes and general 3D road network generation and robot elevator navigation schemes. These schemes only construct a single-layer 2D planar road network. Elevators and stairs between floors are roughly abstracted using fixed time parameters, without independently creating manageable nodes and passage edges. At the elevator modeling level, elevators are treated as simple cross-floor passages, without separating the waiting area and the car, making it impossible to quantify elevator queuing and congestion. At the stair modeling level, one-way upward and downward paths are not separated, and there is a lack of passage capacity limiting mechanisms, making it difficult to avoid the risk of two-way collisions on stairs. Furthermore, classic multi-robot path planning frameworks uniformly set fixed edge passage weights, assuming all robots have completely identical motion performance, ignoring the differentiated driving characteristics of heterogeneous robots. This results in underestimating the path travel time, insufficient subsequent scheduling time, and frequent timing conflicts. In addition, the existing solution does not have complete resource capacity parameters, and cannot quantitatively constrain the number of robots that can be accommodated in corridors, staircases, elevator lobbies and cars at the same time. When multiple robots work concurrently, resource contention conflicts such as corridor congestion, elevator overload, and staircase collisions are very likely to occur, and even robot deadlock may occur, which will greatly reduce the overall operating efficiency of the cluster.

[0004] In summary, existing technologies either focus only on single-machine autonomous mapping and navigation, which cannot adapt to the collaborative scheduling needs of multi-robot clusters; the modeling granularity of vertical passage facilities such as stairs and elevators is coarse, lacking hierarchical virtual resource splitting and passage capacity constraint mechanisms; at the same time, they cannot adapt to the differentiated motion characteristics of heterogeneous robots, making it difficult to output a complete and quantifiable three-dimensional spatiotemporal road network, and unable to support conflict detection and scheduling optimization of large-scale robot clusters.

[0005] 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

[0006] The purpose of this invention is to provide a method, modeling system and computer-readable storage medium for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings. By using elevator dual virtual node layered modeling, staircase unidirectional capacity constraints and heterogeneous robot differentiated time consumption calculation, a three-dimensional spatiotemporal road network with complete resource capacity parameters is constructed, which can accurately identify congestion conflicts of multiple robots in multi-story buildings and support large-scale cluster spatiotemporal scheduling optimization.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings, executed by a robot cluster scheduling server, includes the following steps: S1. Planar topology modeling: Traverse each floor in all floor sets, extract the business anchor points within the floor as planar nodes and store them in the node set, split the bidirectional corridor of the floor into two reverse one-way passage edges and merge them into the edge set; for each one-way passage edge, calculate the differentiated passage time of each heterogeneous robot through the corresponding one-way passage edge according to the planar travel speed of each heterogeneous robot. S2. Vertical staircase passage modeling: Traverse the topology of each staircase segment, extract the up-entry and down-entry entrance nodes, generate one-way up-entry edges and one-way down-entry edges respectively, and incorporate them into the edge set; calculate and generate differentiated passage time corresponding to the two staircase edges based on the low-speed travel parameters of the heterogeneous robot, and uniformly set the edge capacity of all staircase edges. S3. Layered Virtual Modeling of Elevator Composite Resources: Traverse each elevator, create an independent elevator lobby virtual node for each floor of the elevator, and simultaneously create a global elevator car virtual node, and add the lobby virtual node and the global car virtual node to the node set; for any two floors, construct an elevator cross-floor composite passage edge connecting the corresponding floor lobby virtual nodes and add it to the edge set; calculate the heterogeneous robot passage time of the elevator composite passage edge in segments based on the elevator door opening and closing time and the cross-floor operation time between floors; S4. Standardized road network output: The output includes the set of nodes, the three-dimensional spatiotemporal road network map of the one-way travel edge, the differentiated travel time of each side corresponding to each heterogeneous robot, and a complete set of resource capacity parameters for corridors, stairs, and elevators. The three-dimensional spatiotemporal road network map, the differentiated travel time, and the resource capacity parameters are used for spatiotemporal conflict detection and task scheduling of multi-heterogeneous robot clusters.

[0008] Furthermore, following any one or a combination of the aforementioned technical solutions, in step S3, the formula for calculating the travel time of the elevator composite passage side is as follows: = + +

[0009] Among them, g r The time it takes for an elevator door to open and close once. For the elevator from the floor fi Run to the floor fj The time consumed across layers.

[0010] Furthermore, based on any or a combination of the aforementioned technical solutions, the business anchor point in step S1 is one or more of the following: floor business stop point, material storage and retrieval point, and elevator original lobby point.

[0011] Furthermore, based on any one or more of the aforementioned technical solutions or combinations thereof, the complete set of resource capacity parameters includes one or more of the following: planar edge capacity, planar node capacity, elevator lobby waiting area capacity, and elevator car passenger capacity.

[0012] Furthermore, following any one or a combination of the aforementioned technical solutions, each travel edge in the three-dimensional spatiotemporal road network map is bound to a corresponding independent and differentiated travel time for each of the heterogeneous robots, and the road network map contains four types of independent and controllable resources: planar corridors, stairs, elevator lobbies, and elevator cars, each configured with capacity constraint parameters.

[0013] Furthermore, based on any or a combination of the aforementioned technical solutions, the heterogeneous robot includes robots with the same or different driving speeds and load capacities.

[0014] Furthermore, following any one or a combination of the aforementioned technical solutions, step S2 involves setting the staircase side capacity. =1, used to restrict only one robot to use the stairs at a time, prohibiting two-way collisions.

[0015] According to another aspect of the present invention, the present invention provides a three-dimensional spatiotemporal road network modeling system for heterogeneous robots in multi-story buildings, including a robot cluster scheduling server, wherein the scheduling server pre-stores floor sets, stair topology, elevator hardware operating parameters, and motion performance parameters of multiple heterogeneous robots; The scheduling server integrates the following modules that communicate with each other: The planar road network construction module is used to traverse each floor, extract floor business anchor points to generate planar nodes, split the bidirectional corridor into reverse one-way passage edges, and calculate the differentiated passage time corresponding to each heterogeneous robot on each side according to the planar travel speed of each heterogeneous robot. The vertical modeling module for stairs is used to analyze the upper and lower entrance points of each staircase, generate staircase edges that go up one way and down one way respectively, update the staircase passage time at a low speed in conjunction with the heterogeneous robot staircase, and configure a fixed passage capacity for all staircase edges. The elevator composite resource modeling module is used to build an independent elevator lobby virtual node for each floor of each elevator, and a global car virtual node for a single elevator. It generates elevator cross-floor composite edges that connect different floor lobby nodes, and calculates the elevator composite edge travel time in segments based on the elevator's single door opening and closing time and floor cross-floor operation time. The road network output module is used to integrate all nodes and unidirectional edges to generate a three-dimensional spatiotemporal road network map, and synchronously output the differentiated travel time of the heterogeneous robots corresponding to each travel edge, as well as the full range of resource capacity parameters for planes, stairs, and elevators.

[0016] Furthermore, following any one or a combination of the aforementioned technical solutions, the scheduling server connects multiple heterogeneous robots via wired or wireless communication links, and the three-dimensional spatiotemporal road network map output by the scheduling server is sent to the local navigation controller mounted on each of the heterogeneous robots.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when executed by a processor, the computer program implements the method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings 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 using virtual node modeling of the elevator lobby and car at two levels, and setting one-way traffic capacity constraints for the staircase, the queuing and passing congestion of various vertical passages in multi-story buildings can be quantified in a refined manner, solving the problem of rough vertical resource modeling and inability to identify multi-machine resource contention conflicts in the existing solution; b. Calculate the travel time for each heterogeneous robot individually and output a standardized three-dimensional spatiotemporal road network with complete resource capacity parameters. The road network can be directly adapted to discrete time series simulation, providing accurate underlying data support for the global scheduling of large-scale robot clusters. 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 method for constructing a three-dimensional spatiotemporal road network for a heterogeneous robot in a multi-story building, as provided in an exemplary embodiment of the present invention; Figure 2 A schematic diagram of a three-dimensional spatiotemporal road network modeling system for heterogeneous robots in multi-story buildings, 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 invention, see Appendix Figure 1 This paper provides a method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings, which is executed by a robot cluster scheduling server and includes the following steps: S1. Planar topology modeling: Traverse each floor in all floor sets, extract the business anchor points within the floor as planar nodes and store them in the node set, split the bidirectional corridor of the floor into two reverse one-way passage edges and merge them into the edge set; for each one-way passage edge, calculate the differentiated passage time of each heterogeneous robot through the corresponding one-way passage edge according to the planar travel speed of each heterogeneous robot. S2. Vertical staircase passage modeling: Traverse the topology of each staircase segment, extract the up-entry and down-entry entrance nodes, generate one-way up-entry edges and one-way down-entry edges respectively, and incorporate them into the edge set; calculate and generate differentiated passage time corresponding to the two staircase edges based on the low-speed travel parameters of the heterogeneous robot, and uniformly set the edge capacity of all staircase edges. S3. Layered Virtual Modeling of Elevator Composite Resources: Traverse each elevator, create an independent elevator lobby virtual node for each floor of the elevator, and simultaneously create a global elevator car virtual node, and add the lobby virtual node and the global car virtual node to the node set; for any two floors, construct an elevator cross-floor composite passage edge connecting the corresponding floor lobby virtual nodes and add it to the edge set; calculate the heterogeneous robot passage time of the elevator composite passage edge in segments based on the elevator door opening and closing time and the cross-floor operation time between floors; S4. Standardized road network output: The output includes the set of nodes, the three-dimensional spatiotemporal road network map of the one-way travel edge, the differentiated travel time of each side corresponding to each heterogeneous robot, and a complete set of resource capacity parameters for corridors, stairs, and elevators. The three-dimensional spatiotemporal road network map, the differentiated travel time, and the resource capacity parameters are used for spatiotemporal conflict detection and task scheduling of multi-heterogeneous robot clusters.

[0024] This solution constructs three independent edge subsets—plane, staircase, and elevator—in a hierarchical manner and aggregates them into a unified global road network. It distinguishes the topological rules of different passage spaces and calculates the exclusive passage time for each heterogeneous robot, accurately adapting to collaborative scenarios of multiple robot models with different speeds and loads. For elevators, it adds virtual nodes in the lobby and car to complete refined resource modeling. It combines the two-stage time calculation method of elevator door opening and closing and floor segment operation to restore the real elevator riding sequence. At the same time, it uniformly configures fixed capacity constraints for staircases to avoid two-way collisions. Finally, it outputs a standardized 3D spatiotemporal road network with complete classification resource capacity parameters, which can directly support full-dimensional spatiotemporal occupancy verification and conflict detection for large-scale robot clusters. The modeling logic is universal and can be adapted to various multi-story building delivery scenarios such as hospitals, office buildings, and warehouses. It solves the shortcomings of traditional 2D plane road networks, which cannot take into account vertical elevator and staircase resource management, do not distinguish robot performance differences, and have insufficient accuracy in congestion and conflict recognition.

[0025] In one embodiment of the present invention, a method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings is provided. For example, this method can be deployed in a hospital. The following description uses a robot cluster scheduling server deployed in a hospital material distribution scenario as an example. Before executing the road network construction, the server pre-imports basic data related to the multi-story building space, elevators, stairs, and heterogeneous robots to support the complete modeling process. The required basic data includes three categories: building space set, equipment operating parameters, and robot motion parameters. The first category is spatial set data, including the set of floors F consisting of all floors, the set of stair topologies S consisting of all stair structures within the building, and the set of elevators within the building. Simultaneously, a predefined basic storage structure for the road network is used to store the road network node set V, It also pre-sets a set of discrete time steps. The subsequently generated road network can be combined with discrete time steps to complete the temporal and spatial occupancy time sequence deduction of the robot.

[0026] The second category is elevator hardware operating parameters, which apply to each elevator. r Pre-store the time consumed in a single door opening and closing. g r Any floor f 1st floor f Cross-layer operation time of 2 Maximum number of robots that can be accommodated in an elevator car The maximum number of robots that can be accommodated in the waiting area of ​​each elevator lobby. .

[0027] The third category consists of heterogeneous robot motion and traffic capacity parameters, denoted as follows for each robot: Maximum load capacity The required edge and node capacity corresponding to the footprint size are stored separately, and their respective planar travel speeds are stored separately. Staircase low-speed travel; at the same time, uniformly configure the basic capacity of access resources, including the capacity of one side of ordinary corridors. (A value of 1 indicates that only one device is allowed to pass through at a time), floor service node capacity And fix the capacity value of all staircase passage sides to be [value]. =1; Variables generated during modeling. Used to characterize robots The discrete time steps required to travel along a unidirectional edge e; the independent and differentiated travel time for each robot corresponding to each travel edge; global node set. V It uniformly accommodates three types of node elements: floor business anchor points, staircase entrance / exit nodes, and elevator virtual nodes, among which elevator virtual nodes are divided into independent lobby virtual nodes for each floor. Virtual nodes of a single elevator car .

[0028] After the basic data import is completed, the scheduling server completes the construction of the three-dimensional spatiotemporal road network according to the following steps: S1, Planar Layer Topology Modeling Traverse the set of floors F Each floor f Extract floors f Business anchor points such as internal ward locations, pharmacy locations, and original elevator lobby locations are combined to form a unique subset of nodes for this floor. ,Will Merge into the global node set V Identify all bidirectional corridors within a floor, and divide each bidirectional corridor into two unidirectional edges with opposite directions of travel, generating a subset of planar edges. and will The entire set of unidirectional edges is merged into the global set of edges. E For each planar one-way edge in the global set of one-way edges E... e Retrieve each robot Corresponding horizontal travel speed Calculate the differentiated travel time for each robot to travel along the edge. Establish a one-to-one time-consuming mapping relationship between the robot and the passing edge.

[0029] S2, Topology modeling of vertical passageway for stairs Traverse each staircase structure s within the staircase topology set S; extract the upward entrance / exit nodes of that staircase segment. Downlink entrance and exit nodes Generate one-way downward passing edges respectively. = , ), one-way upward passage edge = Generate a subset of stair edges. and will Merge into the global set of one-way edges E .

[0030] Retrieve the robot's low-speed stair-climbing parameters, calculate and update them respectively. , Differentiated passage times for each robot , ; for the staircase edge subset All staircases within the building have a uniform side capacity configuration on one-way sides. =1, this capacity constraint can ensure that only one robot is allowed to occupy the side of the stairs in the same discrete time step, thus avoiding the conflict of two robots meeting each other on the stairs from the bottom structure of the road network.

[0031] S3, Virtual Node Modeling of Elevator Composite Resources Traversing the elevator set R Every elevator inside r Elevator resource modeling was completed in three steps: The first step is to target each floor. Create a virtual lobby node for the floor corresponding to the elevator. At the same time, a unique global virtual car node is created for each elevator r. All lobby virtual nodes and car virtual nodes are merged into the global node set. V ; The second step is to iterate through any set of floor pairs. f i , fj ), construct connecting floors f i Virtual nodes in the lobby and floors fj Elevator cross-floor composite passage edge in the lobby virtual node Generate elevator edge subset and will Merge into the global edge set E。

[0032] The third step involves calculating the robot-specific travel time for each segment of the elevator's composite travel edge, based on the elevator's hardware parameters. The calculation formula is as follows: = + + Among them, the maximum passenger capacity of the elevator car This calculation method fully covers the time loss of the three complete elevator ride process: the robot arriving at the lobby and opening the door, the elevator running between floors, and the robot opening the door again upon arriving at the target floor.

[0033] S4, Standardized Road Network Data Unified Output The scheduling server integrates all node sets. V Global set of unidirectional edges E (Include , , (All subsets), generate a complete 3D spatiotemporal road network map. Simultaneously output three types of standardized structured data: 1) Differentiated travel time matrix corresponding to all accessible edges It fully stores the discrete time steps required for each robot to traverse each access edge; 2) Complete set of resource capacity parameters, including the capacity of the planar corridor side. Floor business node capacity Fixed passage capacity along staircases and capacity of elevator lobby waiting areas Elevator car passenger capacity ; 3) Set of original hardware parameters for the elevator .

[0034] The server outputs road network, time consumption, and capacity parameters with a unified and standardized interface. These parameters can be combined with the discrete time step set T to perform spatiotemporal occupancy simulations for multiple robots. The server also supports generating binary occupancy variables with values ​​of only 0 or 1 for full-scene resource conflict verification. Edge occupancy variable : Marking robots Does the journey begin at time t along the passage edge e? Node occupancy variables : Marking robots Does node v occupy at time t? Elevator lobby occupancy variable : Marking robots At time t, the elevator is in the lobby waiting / entry / exit buffer zone of elevator r; Elevator car occupancy variables : Marking robots It enables resource conflict detection and scheduling calculation across all scenarios, including multi-robot clusters in planar corridors, staircases, elevator lobbies, and elevator cars.

[0035] In the actual implementation of this embodiment, this solution has the following advantages compared with traditional single-layer two-dimensional road network and single-machine grid mapping solutions: First, use independent upward and downward unidirectional edges of the staircase, combined with capacity constraints. =1, which can solidify the stair access rules during the road network modeling stage and accurately identify resource conflicts such as two-way vehicle encounters on stairs; Second, the elevator adopts a two-level virtual resource split model of lobby and car to fully restore two types of vertical congestion scenarios: robot waiting in line and car overload. Based on three-stage time consumption calculation, the complete elevator riding sequence of the robot is accurately restored. Third, calculate the dedicated passage time for each heterogeneous robot separately. Eliminate the path duration prediction deviation caused by the assumption of uniform speed for all robots and improve the accuracy of scheduling timing calculation; Fourth, it provides a complete resource capacity parameter system that can adapt to the spatiotemporal occupancy constraint verification of multiple robots under discrete time steps, and can provide quantitative and complete underlying road network support for the global scheduling of large-scale clusters of multi-state (dozens of) heterogeneous delivery robots in multiple scenarios.

[0036] In one embodiment of the invention, see Appendix Figure 2 This paper presents a 3D spatiotemporal road network modeling system for heterogeneous robots in multi-story buildings. An exemplary system is deployed in a hospital supplies delivery scenario. The core hardware is a robot cluster scheduling server, which can establish data interaction with multiple heterogeneous hospital supplies delivery robots via wired / wireless communication links. Before system operation, the scheduling server persistently stores all the basic data required for modeling locally, including building space sets, elevator hardware parameters, and motion performance parameters of each robot. Based on the stored data, the internal functional modules collaboratively complete the construction of the 3D spatiotemporal road network. During system operation, road network nodes, one-way travel edges, differentiated travel times, and various resource capacity parameters are generated, and standardized road network data is output for use by downstream scheduling equipment.

[0037] The basic data pre-stored locally on the scheduling server is divided into three categories: The first category is building space set data, including the floor set F, staircase topology set S, and elevator set. Meanwhile, a set of storage nodes is reserved. V Set of one-way edges Storage space, predefined set of discrete time steps The first category is used for subsequent simulations of robot spatiotemporal occupancy timing based on the road network; the second category is elevator hardware parameters, storing the time consumed by a single door opening and closing for each elevator r. g r Any floor fi To the floor fj Cross-layer operation time Maximum capacity of the elevator car Maximum capacity of waiting areas on each floor The third category is heterogeneous robots and traffic capacity parameters, for each robot. Maximum load capacity Independent storage plane travel speed Low-speed stair travel, while also configuring planar side capacity. (A value of 1 indicates that only one machine is allowed to pass through a section), Planar node capacity The capacity of all staircases is uniformly limited to a fixed value. =1; Variables generated during system operation Represents robots The number of discrete time steps required for a traversable edge e, and the set of nodes. V Unified to accommodate floor business anchor points, stairwell entrance / exit nodes, and elevator lobby virtual nodes Virtual nodes of elevator cars Four types of nodes.

[0038] The scheduling server integrates four independent functional modules that communicate and work collaboratively with each other. These modules are: a planar road network construction module, a staircase vertical modeling module, an elevator composite resource modeling module, and a road network output module. The specific execution logic of each module is as follows: The planar road network construction module reads the set of floors. F Each floor inside f The corresponding spatial data is used to extract business anchor points such as wards, pharmacies, and original elevator lobbies to generate single-layer node subsets. And merge into the global node set V Identify all bidirectional corridors within a floor, decompose each bidirectional corridor into two opposing unidirectional edges, and merge them into the global edge set. E For each unidirectional edge e in the plane, retrieve the data from each robot. Plane travel speed Calculate the time required for each heterogeneous robot to travel along its dedicated plane edge by edge. Establish a time-consuming mapping relationship between the robot and the passing edge, and cache the calculated differentiated time consumption to the server memory.

[0039] The staircase vertical modeling module reads the staircase topology set. S Extract the location information of each staircase structure s and extract the entrance and exit nodes of the staircase going up. Downlink entrance and exit nodes Generate one-way uplink edges respectively. One-way downward passage edge And add it to the global one-way edge set. E ; retrieve the robot's low-speed parameters on the stairs, and calculate and update the differentiated passage time for each robot on each of the two stair edges. , Write capacity parameters uniformly to all one-way sides of the staircases. =1, which constrains the system at the road network level to allow only a single robot to pass through the stairs at the same discrete time step, thus avoiding two-way collisions.

[0040] Elevator composite resource modeling module, traversing the elevator set R Each elevator r in the system first serves each floor. f Create a corresponding virtual node for the elevator lobby. At the same time, a unique global virtual car node is generated for elevator r. Store all two types of elevator virtual nodes into a node set. V Then iterate through any floor and pair them ( fi , fj Construct an elevator cross-floor composite passage edge connecting the virtual nodes of the two-story lobby. And merge it into the global one-way edge set. E Finally, the time taken to travel along the elevator's composite side is calculated using a piecewise summation formula: = + + It fully covers the elevator travel time for three stages: door opening, cross-floor operation, and secondary door opening, and caches and stores the elevator travel time corresponding to each robot.

[0041] The road network output module reads the complete set of nodes within the server. V Global one-way edge set E Integrate and generate a complete three-dimensional spatiotemporal road network map Simultaneously, three types of standardized data are aggregated and output: first, a robot-differentiated travel time matrix corresponding to all travel edges; second, a matrix of robot-differentiated travel times. Secondly, a complete set of resource capacity parameters, including planar edge capacity, planar node capacity, staircase edge capacity, and elevator lobby capacity. Elevator car capacity Thirdly, the original hardware parameters of the elevator. The road network structured data output by the module can be sent to downstream multi-robot scheduling equipment. Combined with the discrete time step set T, it can perform robot spatiotemporal occupancy simulation and generate edge occupancy variables. Node occupancy variables Elevator lobby occupancy variables Elevator car occupancy variables It supports full-dimensional resource conflict detection and calculation.

[0042] In this embodiment, when the modeling system is working, the scheduling server sends the 3D spatiotemporal road network data generated by the road network output module to the local navigation controllers on each heterogeneous hospital material delivery robot via a communication link. The robot completes local path calculation based on the received road network data. At the same time, the entire system relies on the collaborative design of layered virtual elevator nodes, staircase unidirectional capacity constraints, and robot differentiated time consumption calculation to solve the defects of traditional single-machine mapping equipment that cannot support multi-machine cluster congestion identification, coarse granularity of vertical resource modeling, and large deviation in travel time prediction. It can stably support the road network underlying modeling needs in scenarios where dozens of heterogeneous delivery robots are scheduled concurrently.

[0043] In one embodiment of the present invention, a computer-readable storage medium is provided, which includes any medium with storage function such as a USB flash drive, solid-state drive, optical disk, or read-only memory, and stores a computer program in the medium.

[0044] When the processor of the robot cluster scheduling server reads and runs the computer program, it executes all the steps of the three-dimensional spatiotemporal road network construction method for heterogeneous robots in multi-story buildings described in the aforementioned embodiments, completes the three-dimensional spatiotemporal road network modeling, and outputs standardized road network, differentiated passage time, and various resource capacity parameters.

[0045] This storage medium can be mounted on any robot scheduling hardware device, enabling automatic road network construction without the need for dedicated modeling software. Relying on the program's built-in elevator layered virtual node modeling, staircase unidirectional capacity constraints, and heterogeneous robot differentiated time consumption calculation logic, it outputs complete road network data adapted to multi-robot cluster conflict detection.

[0046] 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.

[0047] 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 constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings, characterized in that, Executed by the robot cluster scheduling server, the process includes the following steps: S1. Planar topology modeling: Traverse each floor in all floor sets, extract the business anchor points within the floor as planar nodes and store them in the node set, split the bidirectional corridor of the floor into two reverse one-way passage edges and merge them into the edge set; for each one-way passage edge, calculate the differentiated passage time of each heterogeneous robot through the corresponding one-way passage edge according to the planar travel speed of each heterogeneous robot. S2. Vertical staircase passage modeling: Traverse the topology of each staircase segment, extract the up-entry and down-entry entrance nodes, generate one-way up-entry edges and one-way down-entry edges respectively and add them to the edge set; calculate the differentiated passage time corresponding to the two staircase edges based on the low-speed travel parameters of the heterogeneous robot, and uniformly set the edge capacity of all staircase edges. S3. Layered Virtual Modeling of Elevator Composite Resources: Traverse each elevator, create an independent elevator lobby virtual node for each floor of the elevator, and simultaneously create a global elevator car virtual node, and add the lobby virtual node and the global car virtual node to the node set; for any two floors, construct an elevator cross-floor composite passage edge connecting the corresponding floor lobby virtual nodes and add it to the edge set; calculate the heterogeneous robot passage time of the elevator composite passage edge in segments based on the elevator door opening and closing time and the cross-floor operation time between floors; S4. Standardized road network output: The output includes the set of nodes, the three-dimensional spatiotemporal road network map of the one-way travel edge, the differentiated travel time of each side corresponding to each heterogeneous robot, and a complete set of resource capacity parameters for corridors, stairs, and elevators. The three-dimensional spatiotemporal road network map, the differentiated travel time, and the resource capacity parameters are used for spatiotemporal conflict detection and task scheduling of multi-heterogeneous robot clusters.

2. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, In step S3, the formula for calculating the travel time of the elevator composite passage edge is: = + + Among them, g r The time it takes for an elevator door to open and close once. For the elevator from the floor fi Run to the floor fj The time consumed across layers.

3. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, The business anchor point in step S1 is one or more of the following: floor business stop point, material storage and retrieval point, and elevator original lobby point.

4. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, The complete set of resource capacity parameters includes one or more of the following: planar edge capacity, planar node capacity, elevator lobby waiting area capacity, and elevator car passenger capacity.

5. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, Each travel edge in the three-dimensional spatiotemporal road network diagram is bound to a corresponding independent and differentiated travel time for each heterogeneous robot. The three-dimensional spatiotemporal road network diagram contains four types of independent and controllable resources: planar corridors, stairs, elevator lobbies, and elevator cars, each configured with capacity constraint parameters.

6. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, The heterogeneous robots include robots with the same or different driving speeds and load capacities.

7. The method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings according to claim 1, characterized in that, Set the staircase side capacity in step S2 =1, used to restrict only one robot to use the stairs at a time, prohibiting two-way collisions.

8. A three-dimensional spatiotemporal road network modeling system for heterogeneous robots in multi-story buildings, characterized in that, This includes a robot cluster scheduling server, which pre-stores floor sets, staircase topology, elevator hardware operating parameters, and motion performance parameters of multiple heterogeneous robots. The scheduling server integrates the following modules that communicate with each other: The planar road network construction module is used to traverse each floor, extract floor business anchor points to generate planar nodes, split the bidirectional corridor into reverse one-way passage edges, and calculate the differentiated passage time corresponding to each heterogeneous robot on each side according to the planar travel speed of each heterogeneous robot. The vertical modeling module for stairs is used to analyze the upper and lower entrance points of each staircase, generate staircase edges that go up one way and down one way respectively, update the staircase passage time at a low speed in conjunction with the heterogeneous robot staircase, and configure a fixed passage capacity for all staircase edges. The elevator composite resource modeling module is used to build an independent elevator lobby virtual node for each floor of each elevator, and a global car virtual node for a single elevator. It generates elevator cross-floor composite edges that connect different floor lobby nodes, and calculates the elevator composite edge travel time in segments based on the elevator's single door opening and closing time and floor cross-floor operation time. The road network output module is used to integrate all nodes and unidirectional edges to generate a three-dimensional spatiotemporal road network map, and synchronously output the differentiated travel time of the heterogeneous robots corresponding to each travel edge, as well as the full range of resource capacity parameters for planes, stairs, and elevators.

9. The multi-story building heterogeneous robot three-dimensional spatiotemporal road network modeling system according to claim 8, characterized in that, The scheduling server connects multiple heterogeneous robots via wired or wireless communication links, and the three-dimensional spatiotemporal road network map output by the scheduling server is sent to the local navigation controllers mounted on each of the heterogeneous robots.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for constructing a three-dimensional spatiotemporal road network for heterogeneous robots in multi-story buildings as described in any one of claims 1-7.