Cross-floor AGV scheduling method and system

Abstract

The invention discloses a cross-floor AGV scheduling method, and the method comprises the steps: constructing a multi-floor integrated environment map: controlling an AGV to traverse an operation region of each floor so as to collect environment feature information, constructing an environment map of each floor, and carrying out the integration; a step of configuring sites and paths: in the multi-floor integrated environment map, setting operation sites, transfer sites and uniformly named cross-floor shared SM transfer sites for each floor, and planning driving paths in each floor and cross-floor paths for transferring through the SM transfer sites; and a scheduling and execution step: editing and issuing a running script containing a cross-floor task instruction to the AGV through the scheduling system, controlling the AGV to move to an SM transfer station of a target floor based on the multi-floor integrated environment map, the SM transfer station and the planned path, and controlling the cross-floor transmission unit to complete floor switching through the scheduling system so as to execute a cross-floor task. And finally, the purpose of improving the cross-floor AGV scheduling intelligent level and the operation efficiency is achieved.

Description

Technical Field

[0001] This invention relates to the field of automated logistics and intelligent manufacturing technology, and more specifically, to a method and system for scheduling AGVs across floors. Background Technology

[0002] With the widespread application of automation technology in warehousing, manufacturing, and other industries, the operating scenarios of AGVs are extending from single-level planes to complex multi-level environments. However, existing scheduling solutions face significant challenges in multi-level scenarios. On the one hand, map building often relies on manual labor or general-purpose tools, resulting in limited accuracy and difficulty in effectively integrating and managing multi-level maps. This leads to independent station settings on each floor, a lack of shared transit benchmarks, and chaotic cross-level path planning, affecting AGV collaborative efficiency.

[0003] On the other hand, existing scheduling systems have fixed functions, rely heavily on preset programs, and cannot flexibly edit task logic. AGV switching between floors often requires manual intervention, resulting in insufficient automation. At the same time, when multiple AGVs work together, cross-floor transport units such as elevators become key bottlenecks, easily leading to resource competition, causing multiple AGVs to wait in a concentrated manner, and reducing overall handling efficiency.

[0004] Furthermore, existing solutions have weak adaptability to dynamic environmental changes, lag in handling path conflicts, and lack real-time monitoring and self-healing mechanisms for equipment status, resulting in room for improvement in system reliability and intelligence. Therefore, there is an urgent need for an AGV scheduling solution capable of achieving accurate multi-floor map integration, intelligent task scheduling, and resource collaboration. Summary of the Invention

[0005] In view of the above-mentioned technical problems in related technologies, the present invention proposes a cross-floor AGV scheduling method and system, which can overcome the above-mentioned shortcomings of the prior art.

[0006] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows: A method for scheduling AGVs across floors; The cross-floor AGV scheduling method includes the following steps: The steps to construct a multi-floor integrated environmental map are: control the AGV to traverse the work areas of each floor to collect environmental feature information, construct environmental maps of each floor and integrate them; The steps for configuring stations and routes are as follows: In the multi-floor integrated environment map, set up work stations, transfer stations and uniformly named cross-floor shared SM transfer stations for each floor, and plan the driving routes within each floor and the cross-floor routes that transfer via the SM transfer stations. The scheduling and execution steps are as follows: The scheduling system edits and issues a running script containing cross-floor task instructions to the AGV, controls the AGV to move to the SM transfer station on the target floor based on the multi-floor integrated environment map, the SM transfer station and the planned path, and controls the cross-floor transmission unit to complete the floor switching to execute the cross-floor task.

[0007] Furthermore, the steps for constructing a multi-floor integrated environmental map specifically include: using SLAM technology, collecting environmental feature data through the lidar and / or visual camera mounted on the AGV to generate maps of each floor; integrating and associating the generated maps of each floor to form the multi-floor integrated environmental map.

[0008] Furthermore, the scheduling system supports visual drag-and-drop editing to customize the AGV's driving parameters, SM transfer station stopping logic, cross-layer transmission unit linkage instructions, and task execution flow, thereby generating the running script.

[0009] Furthermore, the scheduling and execution steps also include dynamic scheduling and optimization steps: when multiple AGVs synchronously request to use the cross-layer transmission unit, the scheduling system dynamically allocates usage rights based on the urgency of the task and the status data of the AGV equipment; monitors the movement status of each AGV in real time, predicts and dynamically avoids path conflicts; monitors the key operating indicators of the AGV in real time, and automatically triggers self-healing actions when the indicators are abnormal.

[0010] Furthermore, the AGV device status data includes battery power and remaining range; the self-healing action includes at least one of adjusting AGV navigation parameters, pausing the current task and having a standby AGV take over the task.

[0011] According to another aspect of the present invention, a cross-floor AGV scheduling system is provided; This multi-floor AGV scheduling system includes: The AGV itself is used to move within the work area and perform tasks; The mapping module is communicatively connected to the AGV body and is used to control the AGV body to traverse each floor to collect environmental feature information, and to construct and integrate a multi-floor integrated environmental map. The site path configuration module is used to set up work sites, transfer sites and uniformly named cross-floor shared SM transfer sites on each floor in the multi-floor integrated environment map, and to plan the driving routes within each floor and the cross-floor routes that transfer via the SM transfer sites. The scheduling system is communicatively connected to the AGV body, the mapping module, and the station path configuration module, respectively. It is used to edit and issue running scripts containing cross-floor task instructions to the AGV body, and control the movement of the AGV body based on the map, station, and path information. The cross-layer transmission unit is communicatively connected to the scheduling system and is used to receive instructions from the scheduling system and cooperate with the AGV body to complete floor switching.

[0012] Furthermore, the mapping module constructs a map based on SLAM technology and is configured to integrate the maps of each floor into the multi-floor integrated environment map.

[0013] Furthermore, the scheduling system includes a script editing unit and a task management unit; the script editing unit supports visual drag-and-drop editing and is used to generate the running script; the task management unit is used to send the running script to the AGV body and manage the task execution status; the scheduling system is also configured with a task optimization engine, which is used to provide optimization suggestions for editing the task script and / or dynamically adjust the script parameters during task execution.

[0014] Furthermore, it also includes a positioning module, used to acquire the position information of the AGV body in real time and feed it back to the scheduling system; the positioning module adopts multimodal positioning fusion technology, which integrates UWB positioning data, IMU inertial measurement data and / or visual SLAM positioning data.

[0015] Furthermore, the scheduling system is also configured to perform dynamic scheduling and optimization, including: when multiple AGVs synchronously request to use the cross-layer transmission unit, dynamically allocating usage rights according to the urgency of the task and the status data of the AGV equipment; monitoring the movement status of AGVs in real time to predict and avoid path conflicts; monitoring AGV operating indicators and triggering self-healing actions when abnormalities occur.

[0016] The beneficial effects of this invention are as follows: By realizing the integrated construction and unified site configuration of multi-floor environmental maps, and integrating a visual script editing and intelligent task scheduling mechanism, the cross-floor path planning of AGVs is made clearer and the degree of automation of task execution processes is significantly improved; Furthermore, by introducing dynamic priority allocation and multi-dimensional optimization strategies, the competitive conflict of key resources such as elevators is effectively alleviated, and the efficiency of multi-AGV collaborative operation and the overall system response capability are improved; At the same time, combined with dynamic environmental perception and anomaly self-healing handling mechanisms, the adaptability and operational reliability of the system in complex scenarios are enhanced, ultimately achieving the goal of comprehensively improving the intelligent level and operational efficiency of cross-floor AGV scheduling. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0018] A cross-floor AGV scheduling method according to an embodiment of the present invention includes the following steps: The steps to construct a multi-floor integrated environmental map are: control the AGV to traverse the work areas of each floor to collect environmental feature information, construct environmental maps of each floor and integrate them; The steps for configuring stations and routes are as follows: In the multi-floor integrated environment map, set up work stations, transfer stations and uniformly named cross-floor shared SM transfer stations for each floor, and plan the driving routes within each floor and the cross-floor routes that transfer via the SM transfer stations. The scheduling and execution steps are as follows: The scheduling system edits and issues a running script containing cross-floor task instructions to the AGV, controls the AGV to move to the SM transfer station on the target floor based on the multi-floor integrated environment map, the SM transfer station and the planned path, and controls the cross-floor transmission unit to complete the floor switching to execute the cross-floor task.

[0019] According to an embodiment of the present invention, a cross-floor AGV scheduling method is provided. In a specific embodiment, the step of constructing a multi-floor integrated environmental map specifically includes: using SLAM technology, collecting environmental feature data through the lidar and / or vision camera mounted on the AGV to generate maps of each floor; integrating and associating the generated maps of each floor to form the multi-floor integrated environmental map.

[0020] According to an embodiment of the present invention, a cross-floor AGV scheduling method is provided. In a specific embodiment, the scheduling system supports visual drag-and-drop editing to customize the AGV's driving parameters, SM transfer station stopping logic, cross-floor transmission unit linkage instructions, and task execution flow, thereby generating the running script.

[0021] According to an embodiment of the present invention, a cross-floor AGV scheduling method is provided. In a specific embodiment, the scheduling and execution steps further include dynamic scheduling and optimization steps: when multiple AGVs synchronously request to use the cross-floor transmission unit, the scheduling system dynamically allocates usage rights based on the urgency of the task and the AGV equipment status data; it monitors the movement status of each AGV in real time, predicts and dynamically avoids path conflicts; and it monitors the key operating indicators of the AGVs in real time, automatically triggering self-healing actions when the indicators are abnormal.

[0022] According to an embodiment of the present invention, a cross-floor AGV scheduling method is provided. In a specific embodiment, the AGV device status data includes battery power and remaining range; the self-healing action includes at least one of adjusting AGV navigation parameters, pausing the current task and having a standby AGV take over the task.

[0023] Secondly, according to an embodiment of the present invention, a cross-floor AGV scheduling system includes: The AGV itself is used to move within the work area and perform tasks; The mapping module is communicatively connected to the AGV body and is used to control the AGV body to traverse each floor to collect environmental feature information, and to construct and integrate a multi-floor integrated environmental map. The site path configuration module is used to set up work sites, transfer sites and uniformly named cross-floor shared SM transfer sites on each floor in the multi-floor integrated environment map, and to plan the driving routes within each floor and the cross-floor routes that transfer via the SM transfer sites. The scheduling system is communicatively connected to the AGV body, the mapping module, and the station path configuration module, respectively. It is used to edit and issue running scripts containing cross-floor task instructions to the AGV body, and control the movement of the AGV body based on the map, station, and path information. The cross-layer transmission unit is communicatively connected to the scheduling system and is used to receive instructions from the scheduling system and cooperate with the AGV body to complete floor switching.

[0024] According to an embodiment of the present invention, a cross-floor AGV scheduling system is provided. In a specific embodiment, the mapping module constructs a map based on SLAM technology and is configured to integrate the maps of each floor into the multi-floor integrated environment map.

[0025] According to an embodiment of the present invention, a cross-floor AGV scheduling system includes, in a specific embodiment, a script editing unit and a task management unit; the script editing unit supports visual drag-and-drop editing and is used to generate the running script; the task management unit is used to distribute the running script to the AGV body and manage the task execution status; the scheduling system is also configured with a task optimization engine, which is used to provide optimization suggestions for editing the task script and / or dynamically adjust the script parameters during task execution.

[0026] According to an embodiment of the present invention, a cross-floor AGV scheduling system further includes a positioning module in a specific embodiment, which is used to acquire the position information of the AGV body in real time and feed it back to the scheduling system; the positioning module adopts multimodal positioning fusion technology to fuse UWB positioning data, IMU inertial measurement data and / or visual SLAM positioning data.

[0027] According to an embodiment of the present invention, a cross-floor AGV scheduling system is provided. In a specific embodiment, the scheduling system is further configured to perform dynamic scheduling and optimization, including: when multiple AGVs synchronously request to use the cross-floor transmission unit, dynamically allocating usage rights according to the urgency of the task and the AGV equipment status data; monitoring the AGV movement status in real time to predict and avoid path conflicts; and monitoring AGV operating indicators and triggering self-healing actions when abnormalities occur.

[0028] To facilitate understanding of the above technical solutions of the present invention, the following detailed description of the above technical solutions of the present invention is provided through specific implementation details and principles.

[0029] In practical use, according to the present invention, a cross-floor AGV scheduling method includes an AGV body, an RDS scheduling system, a Roboshop mapping software module, a site path configuration module, and a cross-floor transmission unit.

[0030] The AGV itself possesses autonomous navigation, motion control, and task execution capabilities, used to complete tasks such as cargo handling and inspection. The RDS scheduling system, as the core control unit, has built-in script editing and task management units for editing and running scripts, creating and issuing task instructions, processing map, station, and path data, and realizing the scheduling and control of the AGV itself. The Roboshop mapping software module communicates with the AGV and the RDS scheduling system, controlling the AGV to traverse multi-floor work areas, collect environmental feature information, construct high-precision environmental maps for each floor, and upload the map data to the RDS scheduling system for storage.

[0031] The site path configuration module connects to the RDS scheduling system to set up work sites, transfer sites, and uniformly named SM transfer sites on the environmental map of each floor. Based on the site distribution, it plans the travel paths within each floor and the cross-floor paths via SM sites, synchronizing this information to the RDS scheduling system. Addressing the efficiency bottleneck of multiple AGVs waiting for the same elevator simultaneously, the linkage logic between the cross-floor transmission unit and the RDS scheduling system is upgraded. Through dynamic priority allocation and multi-elevator collaborative prediction, intelligent scheduling of elevator resources is achieved, reducing AGV waiting time and improving the efficiency of cross-floor material transfer. The core logic of dynamic priority allocation is embedded in the linkage instructions of the cross-floor transmission unit, forming a closed-loop scheduling mechanism: when multiple AGVs simultaneously initiate elevator usage requests, the RDS scheduling system will make intelligent decisions based on two key dimensions. On the one hand, it prioritizes tasks based on their urgency; on the other hand, it dynamically allocates elevator usage rights through an algorithm model, combining real-time battery power, remaining range, and other equipment status data of the AGVs. Furthermore, it flexibly adjusts the elevator operation sequence according to real-time conditions such as changes in on-site tasks and fluctuations in equipment status, achieving precise matching and efficient linkage from task to elevator to AGV.

[0032] This optimization further enhances the practical effectiveness of the intelligent cross-layer transmission unit linkage mechanism, upgrading the traditional passive response scheduling to an intelligent mode of proactive prediction and dynamic allocation. It not only completely solves the congestion problem of multiple AGVs waiting for the same elevator, but also ensures the timely achievement of core tasks through priority-differentiated management, comprehensively improving the overall efficiency and scheduling flexibility of cross-layer material handling.

[0033] Specifically, Roboshop's mapping software module is based on SLAM technology. It uses the LiDAR and vision cameras mounted on the AGV to collect environmental feature data, automatically generating grid maps or semantic maps for each floor, and supporting the integration and annotation of multi-floor maps. Within the existing SLAM mapping technology framework, a dynamic environment adaptation mechanism is introduced and built, including the following key technology upgrades: 1. Intelligent recognition and tracking of dynamic obstacles By fully leveraging the high-precision distance measurement of LiDAR and the rich semantic information of visual cameras, and through data fusion algorithms, the detection rate and classification accuracy of dynamic obstacles are improved. Machine learning / deep learning models are introduced, integrating advanced target detection and tracking algorithms to process sensor data streams in real time. This enables not only the identification of static backgrounds but also the accurate identification and continuous tracking of moving objects.

[0034] 2. Dynamic Map Updates and Representation Layered map management: The map logic is divided into static and dynamic layers.

[0035] Static layer: Constructed by traditional SLAM algorithm, it includes long-term stable environmental features such as walls, pillars, and fixed shelves.

[0036] Dynamic layer: Receives information from the dynamic obstacle recognition module in real time, overlays and displays the position, outline, speed, and direction of movement of dynamic obstacles on the static map, and updates in real time according to their movement.

[0037] The system not only records the current location of dynamic obstacles, but also predicts the areas where they may appear in the future based on their movement trends, and marks them on the map accordingly, providing a more comprehensive basis for path planning.

[0038] 3. Dynamic obstacle-aware path planning and navigation The path planner has been upgraded to be able to perceive and utilize dynamic layer information. When planning a path, it should not only consider static obstacles but also actively avoid dynamic obstacles and their predicted movement areas. During operation, the AGV can dynamically adjust its speed and path based on a real-time updated dynamic map to safely and efficiently avoid sudden or moving obstacles, ensuring smooth and safe navigation.

[0039] The RDS scheduling system's script editing unit supports visual drag-and-drop editing, allowing customization of AGV driving parameters on each floor, SM station docking logic, cross-floor transmission unit linkage instructions, and task execution flows, generating standardized, runnable scripts. To achieve these goals, a task optimization engine based on artificial intelligence and operations research is introduced on top of the existing script editing unit. This includes the following key technology upgrades: 1. Build an intelligent recommendation module An intelligent recommendation module is embedded in the script editing unit. When a user creates or edits a task script, this module analyzes historical and real-time data. The system automatically analyzes historical task execution data, the current position and status of the AGV, the busy level of each SM station, the real-time operating status and load of the elevator, and other multi-dimensional information. Based on the analysis results, the module can provide users with dynamic and optimized path suggestions, station access order suggestions, and scheduling instruction suggestions for cross-floor transmission units.

[0040] 2. Implement dynamic script parameter adjustment The optimization engine not only provides suggestions but also makes dynamic adjustments during script execution. During task execution, the system continuously monitors the global status, and when significant changes are detected, the optimization engine performs a real-time evaluation. Based on the evaluation results, the system can automatically adjust the parameters of the executing script, such as: replanning a better path for the AGV; dynamically adjusting the execution priority of tasks within permissible limits; and intelligently adjusting the request time and order of multiple AGVs for resources such as elevators to avoid concentrated waiting and achieve peak shaving and valley filling.

[0041] 3. Global resource scheduling and conflict resolution To address the issue of severe resource contention often caused by elevator scheduling demands, the optimization solution is as follows: Centralized resource management: The optimization engine centrally manages and schedules all shared resource usage requests. Conflict prediction and resolution: Through a predictive model, potential resource conflicts are identified in advance and resolved by adjusting task execution order, allocating different resources, or suggesting waiting times.

[0042] By introducing an AI-based task optimization engine, the script editing unit of the RDS scheduling system will achieve the following value enhancements: A leap from manual planning to intelligent decision-making: The script editing unit will be upgraded from a simple editing tool to an intelligent decision support system with autonomous analysis, prediction, and optimization capabilities. Alleviating resource contention: Through global optimization and dynamic scheduling, the problem of severe resource contention often caused by elevator scheduling demands is effectively solved, reducing AGV's ineffective waiting time. Improving overall efficiency: By optimizing paths and resource utilization, task completion time is expected to be significantly shortened, and the overall system throughput and response speed are improved. Enhancing system robustness and adaptability: The system can better cope with dynamically changing environments and emergencies, reducing manual intervention and improving automation levels and operational stability. Empowering users and improving experience: Providing users with intelligent assistance reduces the complexity and difficulty of script writing, improving user experience and work efficiency.

[0043] In summary, this optimization suggestion, by introducing advanced AI optimization algorithms, will significantly enhance the intelligent decision-making capabilities of the RDS scheduling system's script editing unit, effectively solve resource competition problems, and greatly improve the overall operating efficiency, adaptability, and intelligence level of the AGV system.

[0044] The locations of the SM stations on each floor precisely correspond to the entrances and exits of the cross-floor transfer units. The coordinate information of the SM stations on each floor is uniformly calibrated in the RDS scheduling system, forming a reference coordinate system for cross-floor transfers. The cross-floor transfer unit is one or more combinations of intelligent elevators, lifting platforms, or chain conveyors. It can receive start / stop, positioning, and floor switching commands from the RDS scheduling system and work with the AGV to complete cross-floor transfers.

[0045] The UWB positioning unit on the AGV body and the positioning beacons deployed on each floor are used to collect the AGV body's position information in real time and feed it back to the RDS scheduling system, so as to realize real-time monitoring of the AGV's movement status and dynamic adjustment of the path.

[0046] To achieve the above objectives, multimodal positioning fusion technology is introduced on the basis of existing UWB positioning, intelligently fusing UWB data with data from various sensors such as IMU and visual SLAM. Specifically, this includes the following key technology upgrades: 1. Multi-sensor data acquisition Add an IMU and a vision camera to the AGV body: UWB: Provides absolute location information on a global scale, serving as the basis for positioning.

[0047] IMU: It can provide high-frequency angular velocity and acceleration data, and can provide high-precision relative position and attitude estimation through trajectory extrapolation in a short time, ensuring short-term continuity of positioning.

[0048] Visual SLAM: Utilizing image information captured by a camera, and through algorithms such as feature point matching, it can not only provide position and pose estimation, but also construct an environmental map, match it with a known map, and achieve localization correction, especially in structured environments where it can provide high accuracy.

[0049] 2. Introduce advanced data fusion algorithms Advanced data fusion algorithms based on Kalman filtering, extended Kalman filtering, or unscented Kalman filtering are added to the positioning module. These algorithms receive positioning data in real time from various sensors, including UWB, IMU, and visual SLAM. The algorithms intelligently utilize the strengths of each sensor, modeling and suppressing their respective measurement errors to output a smoother, more accurate, and more reliable fused position and attitude estimate.

[0050] By introducing multimodal positioning fusion technology, the positioning module will achieve the following value enhancements: the fused positioning accuracy is expected to be significantly better than single UWB positioning, especially in complex environments and during dynamic movements. The system has a higher tolerance for sensor failure, signal obstruction, or interference, and the positioning results are more stable and reliable. It effectively solves the problem of positioning blind spots or weak areas in areas such as the entrances and exits of cross-layer transmission units, ensuring the positioning continuity of the AGV throughout the entire process.

[0051] By providing high-precision and highly reliable real-time location information, a more stable input is provided for image recognition technologies that rely on this information, thus helping to solve the problems of insufficient accuracy and real-time performance in image recognition technology in existing AGV scheduling schemes. Accurate and reliable positioning is a prerequisite for the RDS scheduling system to perform efficient path planning, task allocation, and multi-AGV collaboration. Improved positioning accuracy will directly promote the optimization of scheduling algorithms, reduce path corrections, collision risks, and task delays caused by inaccurate positioning, thereby comprehensively improving the overall task collaboration efficiency and operational stability of the system.

[0052] In summary, this optimization suggestion, by introducing multimodal positioning fusion technology, will significantly improve the positioning accuracy, robustness, and continuity of AGVs in complex environments, providing stronger technical support for the efficient and stable operation of the entire AGV system.

[0053] In its implementation, the RDS scheduling system edits AGV motion control scripts and cross-floor task scripts through the script editing unit. These scripts are then distributed to the AGV via the task management unit. Combined with multi-floor maps, SM stations, and path information constructed using the Roboshop mapping software module, the system controls the AGV to complete cross-floor movement and task execution via the cross-floor transmission unit. When path conflicts or equipment malfunctions are detected, the RDS scheduling system immediately triggers a script adjustment mechanism to replan the path or suspend task execution, ensuring operational safety. Addressing the issue that existing RDS scheduling systems can only passively respond to path conflicts / equipment malfunctions, the system has been upgraded to a predictive and self-healing dual-mode: it anticipates fault risks through sensor data, automatically triggers backup plans, avoids task interruptions or safety hazards, and directly resolves the pain points of delayed anomaly handling and insufficient system reliability.

[0054] This system achieves fully automated control of AGV cross-floor movement by combining multi-floor precise mapping with scripted task editing and unified SM site configuration of the RDS scheduling system. It eliminates the need for manual intervention in floor switching and path adjustment, significantly improving the efficiency and flexibility of AGV scheduling in multi-floor scenarios. It is suitable for multi-level warehousing, intelligent manufacturing and other fields. Example

[0055] This invention provides a cross-floor AGV scheduling system and method, including an AGV body, an RDS scheduling system, a Roboshop mapping software module, a site path configuration module, and a cross-floor transmission unit.

[0056] The AGV itself is an AMB-150 intelligent handling robot, equipped with sensors such as LiDAR, vision cameras, and UWB positioning modules, possessing autonomous navigation, motion control, and task execution capabilities. The RDS scheduling system uses an industrial-grade computer with a built-in SQL Server database for editing and running scripts, creating and issuing task instructions, and achieving AGV scheduling and management. The Roboshop mapping software module uses SLAM technology to collect environmental feature data from the AGV itself, constructing high-precision environmental maps for each floor with a resolution of 0.1 meters. The site path configuration module sets up work stations, transfer stations, and SM transfer stations on each floor's environmental map; SM stations use uniform rectangular markers.

[0057] During implementation, the Roboshop mapping software module controls the AGV to sequentially enter the work areas on each floor, collect environmental feature data, construct grid maps for each floor, and integrate them into a multi-floor unified map. The station path configuration module plans the travel paths within each floor and the cross-floor paths via SM stations according to work requirements. The RDS scheduling system allows editing of the running script through a visual drag-and-drop editing function, setting the AGV's travel speed on each floor to 1 meter / second and the SM station stopping time to 3 seconds. The cross-floor transmission unit linkage command is set to automatically start the elevator to the target floor when the AGV is detected to have arrived at the SM station, and close the elevator door after the AGV arrives. Example

[0058] Based on Example 1, the RDS scheduling system further sets the cross-layer task execution process through the script editing unit: First, it determines whether there are any pending tasks; if there are no tasks, it waits. If there are tasks, it determines whether cross-layer operation is required based on the task content; if not, it executes the task directly. If cross-layer operation is required, it selects a suitable cross-layer transmission unit to switch floors based on the target floor set in the task. Finally, it feeds back the task execution result to the task management unit. When multiple AGVs request to use the same elevator, the optimization engine will intelligently sort and dynamically allocate elevator usage rights based on factors such as task urgency, the estimated time for the AGVs to arrive at the elevator entrance, and the current position of the elevator, or recommend that some AGVs wait for another elevator that is about to become available. Example

[0059] Building upon Example 2, an error handling mechanism for task execution is further implemented. This mechanism primarily monitors three key indicators of the AGV: Drive system: motor current and wheel speed fluctuations; Navigation system: signal strength and positioning accuracy of LiDAR or vision sensors; Power system: battery temperature and remaining power. An AGV health monitoring script is added to the task management unit of the RDS scheduling system, collecting sensor data every 5 seconds. Thresholds are set for these key indicators. If an indicator exceeds the normal range three times consecutively, a fault warning and self-healing action is initiated. For minor anomalies, navigation parameters are automatically adjusted to maintain task execution; for severe anomalies, the current AGV task is immediately suspended, and the scheduling system automatically assigns a backup AGV to take over the task, while simultaneously generating a maintenance work order and pushing it to the administrator's app. In scenarios involving multiple AGVs working collaboratively, path conflict prediction and dynamic avoidance are automatically implemented. Each AGV is assigned a unique ID, and a real-time AGV trajectory database is established in the RDS system, recording the position, speed, and target path of each device. Based on the real-time position and speed of the AGVs, it is calculated whether there will be path intersection within the next 5 seconds. The AGVs that start later will slow down and wait for the vehicle in front to pass through the detour. If slowing down will cause the task to time out, a temporary detour route will be automatically planned. If one of the AGVs is transporting urgent materials, the low-priority AGVs will automatically give way, reducing the occurrence rate of path conflicts by 80%, avoiding frequent start and stop of AGVs, and improving transportation efficiency.

[0060] In summary, by utilizing the technical solutions described above in this invention, the integrated construction and unified site configuration of multi-floor environmental maps, along with the integration of visual script editing and intelligent task scheduling mechanisms, significantly improve the clarity of AGV cross-floor path planning and the automation level of task execution processes. Furthermore, by introducing dynamic priority allocation and multi-dimensional optimization strategies, the competitive conflicts of key resources such as elevators are effectively alleviated, enhancing the efficiency of multi-AGV collaborative operations and the overall system responsiveness. Simultaneously, the dynamic environmental perception and anomaly self-healing mechanisms enhance the system's adaptability and operational reliability in complex scenarios, ultimately achieving the goal of comprehensively improving the intelligence level and operational efficiency of cross-floor AGV scheduling.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dispatching AGVs across floors, characterized in that, The method comprises the following steps: a step of constructing a multi-floor integrated environment map: controlling the AGV to traverse each floor work area to collect environmental feature information, constructing each floor environment map and integrating; a step of configuring stations and paths: in the multi-floor integrated environment map, setting work stations, transfer stations and uniformly named SM transfer stations for each floor, and planning travel paths in each floor and cross-floor paths via the SM transfer stations; a step of scheduling and executing: through the scheduling system, editing and issuing a running script containing cross-floor task instructions to the AGV, controlling the AGV to move to the SM transfer station of the target floor based on the multi-floor integrated environment map, the SM transfer station and the planned path, and through the scheduling system, controlling the cross-layer transfer unit to complete floor switching to execute the cross-floor task.

2. The AGV scheduling method across floors according to claim 1, wherein, The step of constructing a multi-floor integrated environment map specifically comprises: using SLAM technology, collecting environmental feature data through the laser radar and / or visual camera carried by the AGV to generate each floor map; integrating and associating the generated each floor map to form the multi-floor integrated environment map.

3. The AGV scheduling method across floors according to claim 1, wherein, The scheduling system supports visual drag-and-drop editing to customize the travel parameters of the AGV, the SM transfer station docking logic, the cross-layer transfer unit linkage instructions and the task execution process, thereby generating the running script.

4. The AGV scheduling method across floors according to claim 1, wherein, The step of scheduling and executing further comprises a dynamic scheduling and optimization step: when multiple AGVs simultaneously request to use the cross-layer transfer unit, the scheduling system dynamically allocates the use right according to the task urgency and AGV device state data; real-time monitoring of the motion state of each AGV, predicting and dynamically avoiding path conflicts; real-time monitoring of the key operation indicators of the AGV, and automatically triggering self-healing actions when the indicators are abnormal.

5. The AGV dispatching method across floors according to claim 4, wherein, The AGV device state data includes battery power and remaining endurance; the self-healing action includes at least one of adjusting the AGV navigation parameters, suspending the current task and taking over the task by a standby AGV.

6. A cross-floor AGV dispatching system for performing the cross-floor AGV dispatching method of any one of claims 1-5, wherein, The method comprises: an AGV body for moving in a work area and executing tasks; a mapping module in communication with the AGV body for controlling the AGV body to traverse each floor to collect environmental feature information and construct and integrate a multi-floor integrated environment map; a station and path configuration module for setting work stations, transfer stations and uniformly named cross-floor shared SM transfer stations for each floor in the multi-floor integrated environment map, and planning travel paths in each floor and cross-floor paths via the SM transfer stations; a scheduling system in communication with the AGV body, the mapping module and the station and path configuration module, respectively, for editing and issuing a running script containing cross-floor task instructions to the AGV body, and controlling the motion of the AGV body based on the map, station and path information; a cross-layer transfer unit in communication with the scheduling system for receiving instructions from the scheduling system and cooperating with the AGV body to complete floor switching.

7. The cross-floor AGV dispatching system of claim 6, wherein, The mapping module constructs a map based on a SLAM technology, and is configured to integrate each floor map into the multi-floor integrated environment map.

8. The cross-floor AGV dispatching system of claim 6, wherein, The scheduling system comprises a script editing unit and a task management unit; the script editing unit supports visual drag-and-drop editing for generating the running script; and the task management unit is used for issuing the running script to the AGV body and managing the task execution state. The scheduling system is further configured with a task optimization engine for providing optimization suggestions for editing of the task script and / or dynamically adjusting script parameters during task execution.

9. The cross-floor AGV dispatching system of claim 6, wherein, The scheduling system is further configured with a positioning module for acquiring position information of the AGV body in real time and feeding back to the scheduling system; the positioning module adopts a multi-modal positioning fusion technology to fuse UWB positioning data, IMU inertial measurement data and / or visual SLAM positioning data.

10. The cross-floor AGV dispatching system of claim 6, wherein, The scheduling system is further configured to perform dynamic scheduling and optimization, including: when multiple AGVs synchronously request to use the cross-layer transmission unit, dynamically allocating the use right according to the task urgency and AGV device state data; monitoring the AGV motion state in real time to predict and avoid path conflicts; monitoring AGV operation indexes and triggering self-healing actions when an exception occurs.