AMR cluster task planning cloud platform based on digital twinning

Abstract

The invention relates to the technical field of industrial automation and logistics intellectualization, and discloses an AMR cluster task planning cloud platform based on digital twinning. The platform comprises a task scheduling module, a twin modeling module, a cluster state analysis module, a path planning module, a core operation module and the like. The task scheduling module decomposes job demand data; the twin modeling module generates two task topology models; the cluster state analysis module determines a scheduling level; the path planning module provides navigation parameters; and the core arithmetic unit generates a final scheduling instruction. In addition, an equipment collaboration module, an instruction conversion module, a task filing module and the like are further arranged. The platform can accurately plan tasks, improve the AMR cluster operation efficiency, realize resource optimization configuration, solve space-time contradiction between tasks, record the whole operation process, and provide an effective scheme for AMR cluster task planning in industrial production, logistics storage and other scenes.

Description

Technical Field

[0001] The present invention relates to the field of industrial automation and intelligent logistics technology, and specifically to an AMR cluster task planning cloud platform based on digital twins. Background Art

[0002] In modern industrial production and logistics warehousing, autonomous mobile robots (AMRs), with their high flexibility and autonomy, are becoming key equipment for improving production efficiency and logistics delivery accuracy. As their application scale continues to expand, efficient task planning and scheduling for AMR swarms has become a critical issue that needs to be addressed.

[0003] Traditional AMR task planning methods often rely on manual experience or simple algorithms, which have many limitations when handling complex tasks and large-scale clusters. On the one hand, manual scheduling struggles to cope with rapidly changing production demands and complex task scenarios, and is prone to problems such as irrational task allocation, idle or overloaded equipment, and overall low efficiency. For example, in large e-commerce warehouses, order demand fluctuates constantly. Manual scheduling cannot timely and appropriately arrange AMRs for cargo handling based on factors such as order urgency and cargo storage location, often resulting in cargo backlogs or delivery delays.

[0004] On the other hand, simple algorithms cannot achieve globally optimal task planning when dealing with complex situations such as multi-tasking concurrency, limited resources, and dynamic environmental changes. For example, early fixed-rule path planning algorithms failed to account for dynamic factors such as obstacles and interference from other equipment that AMRs may encounter during operation. This led to frequent path conflicts when AMRs were performing tasks, reducing operational efficiency and potentially causing safety incidents such as equipment collisions.

[0005] Furthermore, traditional task planning systems lack robust monitoring of AMR equipment status and resource management. Accurate and real-time information on equipment energy consumption, remaining battery life, and fault status is not readily available, making it difficult to rationally allocate tasks and schedule resources based on the equipment's actual conditions. This impacts the equipment's lifespan and overall system stability.

[0006] With the rise of digital twin technology, its application in the industrial sector has gradually attracted attention. By constructing virtual models of physical entities, digital twins can reflect the state and behavior of physical entities in real time, providing strong support for optimized decision-making. However, research on the application of digital twin technology to AMR swarm task planning is still in the exploratory stage. Existing systems and methods still have shortcomings in task decomposition, model construction, resource matching, and scheduling optimization. These systems and methods cannot fully utilize the advantages of digital twin technology and cannot meet the needs of modern industry and logistics for efficient task planning for AMR swarms. Therefore, the development of a digital twin-based AMR swarm task planning cloud platform is of great practical significance. Summary of the Invention

[0007] The purpose of the present invention is to provide an AMR cluster task planning cloud platform based on digital twins to solve the problems raised in the above background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an AMR cluster task planning cloud platform based on digital twins, the platform comprising:

[0009] A task scheduling module is used to receive multi-source job demand data and identify task types, and divide the job demand data into a basic task data stream and an extended task data stream based on preset task decomposition rules;

[0010] a twin modeling module, configured to perform three-dimensional spatial gridding processing on the basic task data stream to generate a first task topology model, and perform dynamic resource matching processing on the extended task data stream to generate a second task topology model;

[0011] A cluster status analysis module is used to fuse the first task topology model and the second task topology model according to a preset constraint matrix, and then map them to corresponding scheduling levels, and use the scheduling levels as the task allocation benchmark for the current cluster;

[0012] A path planning module, configured to call an optimal path solution from a preset path database based on the scheduling level, and use the optimal path solution as a navigation parameter for the current cluster;

[0013] The core computing unit is used to perform logic verification on the task allocation benchmark and the navigation parameter to generate a final scheduling instruction.

[0014] Preferably, the twin modeling module performs dynamic resource matching processing on the extended task data flow, including:

[0015] Dividing the task sequence in the extended task data stream into a fixed task group and a flexible task group, and calculating the resource occupancy rate of the flexible task group based on a preset dynamic load balancing algorithm to generate a load feature set;

[0016] Performing hierarchical processing on the device status data in the extended task data stream, extracting the energy consumption characteristics of each layer of equipment and constructing an energy consumption map;

[0017] Dynamically associate and match the load feature set with the energy consumption spectrum to generate the second task topology model.

[0018] Preferably, the preset task decomposition rules include a rigid constraint set and a flexible constraint set; the rigid constraint set includes a time window identifier, an equipment capacity identifier and a job priority identifier; the flexible constraint set includes an energy consumption threshold identifier and a path conflict identifier, and each identifier corresponds to an independent task processing channel.

[0019] Preferably, the platform further includes a data interaction interface, which is used to realize communication docking between the task scheduling module, the twin modeling module, the cluster state analysis module and the path planning module and the physical device network respectively;

[0020] The task scheduling module divides the job demand data based on the preset task decomposition rules, including:

[0021] Acquire a mixed data packet from a physical device network in real time through the data interaction interface, and match the protocol label of the mixed data packet according to the identifier in the rigid constraint set to separate the basic task segments;

[0022] Traversing the extended tags of the mixed data packet according to the identifier in the flexible constraint set to extract the flexible task segment;

[0023] The basic task segment and the flexible task segment are synchronized and aligned according to the task triggering time and then written into the task storage area and the extended task buffer area respectively.

[0024] Preferably, when the preset constraint matrix adopts a static analysis model, the scheduling level is an interval mapping result of the linear superposition value of the first task topology model and the second task topology model;

[0025] When the preset constraint matrix adopts a dynamic adjustment model, the scheduling level is a continuous state set in which the associated data of the first task topology model and the second task topology model are corrected in real time through an incremental learning algorithm.

[0026] Preferably, it further comprises a device collaboration module connected to the core computing unit, wherein the device collaboration module is connected to the physical device resource pool via the data interaction interface;

[0027] The device coordination module is configured to screen available device queues from the physical device resource pool according to the device scheduling requirements in the final scheduling instruction, and generate a task execution sequence to optimize the cluster operation process.

[0028] Preferably, the device collaboration module generates a task execution sequence including:

[0029] Loading a three-dimensional spatial topology model, and locating the spatial coordinate point of each device in the available device queue in the topology model;

[0030] Calculate the optimal movement trajectory from the current position of each device to the target operation area based on a dynamic programming algorithm, and sort the available device queues according to the urgency of the tasks;

[0031] The optimal movement trajectory and the efficiency ranking are integrated into the topology model to generate a visual task execution sequence.

[0032] Preferably, when the core computing unit performs logical verification on the task allocation benchmark and navigation parameters, it adopts a dual verification mode of abnormal state detection mechanism and constraint conflict verification mechanism. The abnormal state detection mechanism is used to confirm the integrity of the equipment operation, and the constraint conflict verification mechanism is used to resolve the time and space contradictions between tasks.

[0033] Preferably, it also includes an instruction conversion module connected to the core computing unit, and the instruction conversion module is used to convert the final scheduling instruction into a device control code, and send the device control code to the designated device cluster through the data interaction interface to start the task program.

[0034] Preferably, it also includes a task archiving module connected to the core computing unit, and the task archiving module is used to store the job requirement data, the first task topology model, the second task topology model, the scheduling level and the final scheduling instruction, and generate a complete job record chain according to the task life cycle.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] In terms of task scheduling, the task scheduling module receives multi-source job demand data and identifies the task type, and divides the job demand data into basic task data streams and extended task data streams according to the preset task decomposition rules. The preset task decomposition rules include a rigid constraint set and a flexible constraint set. The time window identifier, equipment capacity identifier, and job priority identifier in the rigid constraint set, as well as the energy consumption threshold identifier and path conflict identifier in the flexible constraint set, can ensure the rationality and efficiency of task processing. For example, based on the time window identifier, urgent tasks can be given priority to avoid task delays; based on the equipment capacity identifier, tasks can be reasonably allocated to prevent equipment overload. Moreover, through the data interaction interface, mixed data packets are obtained from the physical device network in real time, and the basic task segments and elastic task segments are accurately separated and extracted. They are stored after synchronous alignment according to the task trigger time, ensuring the timeliness and orderliness of task processing.

[0037] The design of the twin modeling module is very critical. The basic task data stream is processed by three-dimensional spatial gridding to generate the first task topology model, which provides an accurate spatial basis for subsequent task allocation and path planning. The extended task data stream is processed by dynamic resource matching to generate the second task topology model. Specifically, the task sequence is divided into fixed task groups and flexible task groups. The resource occupancy rate of the flexible task group is calculated based on the preset dynamic load balancing algorithm to generate a load feature set. At the same time, the equipment status data is hierarchically processed to construct an energy consumption map, and then the two are dynamically associated and matched. This enables task planning to fully consider task requirements and equipment resource status and achieve optimal resource allocation. For example, in a logistics warehouse scenario, tasks are reasonably allocated based on factors such as the urgency and weight of different cargo handling tasks and the power and load capacity of each AMR to avoid some AMRs being overworked while others are idle, thereby improving overall operational efficiency.

[0038] The cluster state analysis module fuses the data of the two task topology models based on a preset constraint matrix and maps them to corresponding scheduling levels. When the preset constraint matrix uses a static analysis model, the scheduling level is the interval mapping result of the linear superposition of the two models. When using a dynamic adjustment model, an incremental learning algorithm is used to modify the associated data in real time to generate a continuous state set. This flexible scheduling level determination method adapts to different task scenarios and dynamically changing environments, providing a scientific and reasonable benchmark for task allocation.

[0039] The path planning module uses the optimal path from a pre-set path database as a navigation parameter based on the dispatch level. This path, determined by combining the actual task and equipment status, effectively avoids path conflicts and improves AMR operational efficiency. In complex production workshops or warehouse environments, AMRs can quickly and accurately reach their destinations using the optimal path, reducing travel time and energy consumption.

[0040] The core computing unit performs logical verification of task assignment benchmarks and navigation parameters, employing a dual verification model consisting of anomaly detection and constraint conflict verification. The anomaly detection mechanism confirms the integrity of equipment operation and promptly identifies potential equipment failures. The constraint conflict verification mechanism resolves temporal and spatial conflicts between tasks, ensuring the safety and reliability of task execution. For example, when multiple AMRs are performing tasks simultaneously, this prevents them from arriving at the same location at the same time, thus preventing collisions.

[0041] The device collaboration module screens available device queues based on the final scheduling instructions and generates a task execution sequence. By loading a 3D spatial topology model, locating device spatial coordinates, and calculating optimal movement trajectories using a dynamic programming algorithm, the module prioritizes tasks by urgency and efficiency, generating a visual task execution sequence. This streamlines collaboration between devices, further optimizing cluster operation processes and improving overall operational efficiency.

[0042] The instruction conversion module converts the final scheduling instructions into device control codes and sends them to the designated device cluster to initiate the task program, ensuring the accuracy of instruction transmission and the effectiveness of device execution. The task archiving module stores job demand data, task topology models, scheduling levels, and final scheduling instructions. It generates a complete job record chain based on the task lifecycle, facilitating subsequent query, analysis, and optimization, and contributing to the continuous improvement of task planning strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a working principle diagram of the AMR cluster task planning cloud platform based on digital twins described in the present invention;

[0044] Figure 2 A diagram showing the working principle of the twin modeling module for processing the extended task data stream;

[0045] Figure 3 A diagram showing the working principle of the task scheduling module that divides data based on preset task decomposition rules;

[0046] Figure 4 Generate a working principle diagram of the task execution sequence for the device collaboration module;

[0047] Figure 5 This is the working principle diagram of the core operation unit logic verification. DETAILED DESCRIPTION

[0048] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] See also Figure 1-Figure 5 The present invention provides an AMR cluster task planning cloud platform based on digital twins, and the specific implementation steps are as follows:

[0050] The task scheduling module receives multi-source job demand data and identifies the task type, and divides the job demand data into a basic task data stream and an extended task data stream based on the preset task decomposition rules. Next, the twin modeling module performs three-dimensional spatial gridding processing on the basic task data stream to generate a first task topology model, and performs dynamic resource matching processing on the extended task data stream to generate a second task topology model. Afterwards, the cluster status analysis module performs data fusion on the first task topology model and the second task topology model according to the preset constraint matrix, and maps them to the corresponding scheduling level, and uses the scheduling level as the task allocation benchmark for the current cluster. Then, the path planning module calls the optimal path solution in the preset path database based on the scheduling level, and uses the optimal path solution as the navigation parameter of the current cluster. Finally, the core computing unit performs logical verification on the task allocation benchmark and the navigation parameters to generate the final scheduling instructions.

[0051] The present invention will be further described below in conjunction with Examples 1 to 5:

[0052] Example 1:

[0053] In this cloud platform, the twin modeling module implements a rigorous process for dynamic resource matching within the extended task data stream. First, the task sequence within the extended task data stream is divided into fixed task groups and flexible task groups. In real-world task processing scenarios, tasks vary in nature. Some tasks have fixed execution modes and requirements, while others can be flexibly allocated based on actual resource availability. This is the rationale for the division into fixed and flexible task groups. After this division, resource utilization is calculated for the flexible task groups based on a pre-set dynamic load balancing algorithm. This algorithm comprehensively considers various resource factors, such as computing and storage resources, and generates a load signature set by analyzing and integrating the resource requirements of each task within the flexible task group. This load signature set comprehensively reflects the resource utilization and changing trends of the flexible task group during execution.

[0054] Device status data in the extended task data stream is processed hierarchically. During device operation, devices at different levels have different functional and performance characteristics, resulting in varying energy consumption. This hierarchical processing allows for a more detailed analysis of the energy consumption characteristics of devices at each level. During this processing, device energy consumption data is collected, organized, and analyzed, extracting data indicators that represent the energy consumption characteristics of devices at that level. This energy consumption map is then constructed. This energy consumption map intuitively displays the energy consumption distribution and changing patterns of devices at different levels.

[0055] The load signature set is dynamically correlated with the energy consumption profile to generate a second mission topology model. The matching process fully considers the relationship between load and energy consumption characteristics to find the optimal matching combination. This dynamic correlation and matching ensures that the generated second mission topology model better aligns with actual mission requirements and resource availability, improving the rationality and efficiency of mission planning.

[0056] In a large e-commerce warehouse's cargo sorting and delivery scenario, a large number of orders are generated daily, requiring a cluster of autonomous mobile robots (AMRs) to collaborate on sorting, handling, and delivering the goods. This task data is continuously fed into the cloud platform's task scheduling module, forming an extended task data stream.

[0057] The cloud platform's twin modeling module begins dynamic resource matching for the extended task data stream. The first step is to divide the tasks into groups. Some warehouse sorting tasks have fixed processes, such as sorting specific categories of goods from fixed shelves to fixed temporary storage areas. These tasks are assigned to fixed task groups. However, tasks like expedited sorting to respond to unexpected orders or replenishment based on real-time inventory adjustments, due to their flexible execution time and resource requirements, are assigned to flexible task groups.

[0058] After task groups are divided, resource utilization is calculated for each elastic task group based on a preset dynamic load balancing algorithm to generate a load signature set. For example, within a certain time period, the elastic task group contains three expedited order sorting tasks. Each task has different AMR usage time, power consumption, and warehouse management system computing resource usage. The preset dynamic load balancing algorithm comprehensively considers these factors by analyzing and integrating the various resources expected to be used by each task during execution. For example, suppose one task is expected to require the AMR to operate continuously for 2 hours, consume 30% of power, and occupy 10% of system computing resources; another task is expected to operate for 1.5 hours, consume 20% of power, and occupy 8% of computing resources. After analyzing and processing this data, the algorithm ultimately generates a load signature set that comprehensively reflects the resource usage of the elastic task group. This signature set includes information such as the resource utilization ratio and time distribution of each task, providing a basis for subsequent resource matching.

[0059] The device status data in the extended task data stream is processed hierarchically, and energy consumption characteristics of each device level are extracted to construct an energy consumption map. Warehouse equipment includes automated mobile robots (AMRs), shelf elevators, and cargo conveyors. Taking AMRs as an example, these levels are divided based on factors such as battery capacity and motor power. AMRs at different levels consume different amounts of energy when performing the same task. For each AMR level, energy consumption data is collected under different task scenarios, such as power consumption when moving the same distance with and without a full load, and power consumption when sorting goods of different weights. After organizing and analyzing this data, key energy consumption characteristics of each AMR level are extracted, such as peak and average energy consumption. These energy consumption characteristics are then presented graphically to create an energy consumption map, clearly demonstrating the energy consumption distribution and changing patterns of AMRs at different levels.

[0060] The load feature set is dynamically associated and matched with the energy consumption map to generate a second task topology model. Combining the previously generated load feature set and energy consumption map, the cloud platform will search for the best matching combination between the two. Suppose a certain expedited sorting task in the load feature set has specific requirements for the AMR's usage time and power consumption, and the energy consumption map shows that the AMRs at a certain level have the lowest energy consumption and can complete the task efficiently while meeting the task's time requirements. The cloud platform will match this task with the AMRs at that level, while taking into account other relevant factors such as the task's priority and the current location of the equipment, ultimately generating a comprehensive and reasonable second task topology model. This model clarifies the matching relationship between each task and the corresponding equipment, as well as the execution order and dependencies between tasks, providing an accurate basis for subsequent task scheduling and execution.

[0061] Example 2:

[0062] Pre-set task decomposition rules play a key role in the entire task planning process. They include a set of rigid and flexible constraints. Rigid constraints include time window identifiers, equipment capacity identifiers, and job priority identifiers. The time window identifier specifies the time range within which a task can be executed, a crucial constraint for ensuring timely completion. The equipment capacity identifier specifies the equipment capacity required to execute the task, preventing task execution failures due to insufficient equipment capacity. The job priority identifier determines the order in which different tasks are prioritized, ensuring that important tasks are processed first.

[0063] The flexible constraint set includes energy threshold flags and path conflict flags, each corresponding to a separate task processing channel. The energy threshold flag sets an upper limit on energy consumption during task execution to achieve energy conservation. The path conflict flag is used to detect and avoid path conflicts during task execution, ensuring safe and efficient task execution.

[0064] In practical applications, when the task scheduling module divides job demand data based on preset task decomposition rules, it obtains mixed data packets in real time from the physical device network through a data interaction interface. The data interaction interface is responsible for implementing communication between the task scheduling module and the physical device network to ensure accurate data transmission. After obtaining the mixed data packet, the protocol tag of the mixed data packet is matched according to the identifiers in the rigid constraint set to separate the basic task segments. The protocol tag contains key information about the task. By matching it with the identifiers in the rigid constraint set, the basic task segments can be accurately identified. Then, the extended tags of the mixed data packet are traversed according to the identifiers in the flexible constraint set to extract the flexible task segments. The extended tags record information related to the flexible constraints, and by traversing these tags, the flexible task segments can be extracted. Finally, the basic task segments and the flexible task segments are synchronized and aligned according to the task trigger time and written to the task storage area and the extended task buffer, respectively, for subsequent processing.

[0065] Take a large manufacturing plant as an example. The plant produces a variety of products and has a large number of operational demands every day, such as raw material handling, parts processing, and finished product assembly. These demands are input into the AMR cluster task planning cloud platform based on digital twins in the form of operational demand data. In this scenario, this embodiment has the following specific implementation methods.

[0066] Factory production operations have strict time constraints, highlighting the importance of time windows within rigid constraint sets. For example, the production of a batch of products requires the transportation of raw materials within a specific timeframe to ensure that subsequent processing can proceed on schedule. Failure to complete the transportation task within the specified timeframe will delay the entire production process and increase production costs. Time windows act like a precise clock, setting strict time limits for task execution.

[0067] Equipment capacity identification is also crucial. Different processing equipment has different capacity limitations during component processing. For example, a large CNC machining center has a limited number and size of parts it can process at a time. When scheduling processing tasks, the equipment capacity identification must be used to determine the processing quantity for each batch to ensure that the equipment does not malfunction or affect processing accuracy due to overload. This is like setting a "load limit" for the equipment, ensuring that the equipment operates safely and efficiently.

[0068] Job prioritization plays a key role in coordinating multiple tasks. When a factory receives an urgent order, the associated production tasks are given higher priority. For example, if a key customer urgently requires a batch of parts with specific specifications, all tasks related to producing these parts, including raw material handling, processing, and shipping of the finished product, will be prioritized over other routine tasks. This ensures that urgent orders can be delivered quickly, meeting customer needs and maintaining the factory's reputation.

[0069] Energy consumption thresholds within flexible constraint sets are crucial for energy conservation and emission reduction in factories. For example, each AMR transport robot in a factory consumes electricity during operation. To reduce overall energy consumption, the cloud platform sets energy consumption thresholds. For example, a certain model of AMR must not consume more than a certain amount of energy when completing a transport task. If, during task planning, a transport solution is identified that may cause the AMR's energy consumption to exceed the threshold, the cloud platform will replan the path or adjust the transport strategy to select a more energy-efficient solution, thereby achieving energy savings.

[0070] Path conflict identification is key to ensuring smooth logistics within the factory. In the factory's production workshop, AMR robots need to move within a limited space when transporting goods. Without proper planning, path conflicts between robots can easily occur, leading to traffic jams and affecting production efficiency. Path conflict identification is like the "traffic lights" in traffic regulations. By analyzing and planning the travel paths of each task, it prevents different AMRs from entering the same area at the same time. For example, when two AMRs need to pass through a narrow passage, the cloud platform will reasonably arrange their passage order based on the path conflict identification to ensure safe and efficient logistics transportation.

[0071] The data exchange interface plays a crucial role in the task scheduling module's division of job demand data based on pre-set task decomposition rules. Various factory devices, such as production equipment, handling robots, and sensors, exchange data with the cloud platform via a physical device network. The task scheduling module uses the data exchange interface to receive mixed data packets containing various job demand information in real time.

[0072] The protocol tags of mixed data packets are matched according to the identifiers in the rigid constraint set to separate the basic task segments. For example, a mixed data packet contains information on various tasks, such as raw material handling, parts processing, and finished product assembly. By matching with rigid constraints such as time window identifiers, equipment capacity identifiers, and job priority identifiers, basic tasks with fixed time requirements, clear equipment requirements, and higher priorities can be accurately identified. For example, according to the production plan, a production line needs to replenish raw materials at a fixed time period every day. This task will be identified as a basic task segment and separated from the mixed data packet.

[0073] The extended tags of the mixed data packets are traversed according to the identifiers in the flexible constraint set to extract the elastic task segments. The extended tags record information related to flexible constraints such as energy consumption and path planning. For example, for some non-urgent transport tasks, their execution time and path can be adjusted according to actual conditions. By traversing the extended tags and combining the energy consumption threshold identifier and the path conflict identifier, these elastic task segments can be extracted. For example, if a certain transport task can choose a route with lower energy consumption but a slightly longer path without affecting the overall production progress, this task will be extracted as an elastic task segment.

[0074] The basic task segments and flexible task segments are synchronized and aligned according to the task trigger time, and then written into the task storage area and extended task buffer respectively. During the factory production process, the trigger time of each task varies. For example, the raw material handling task may be triggered when the raw materials arrive, while the parts processing task is triggered after the raw materials are handled. The cloud platform will synchronize and align the separated basic task segments and flexible task segments according to the task trigger time to ensure that the execution order of the tasks is reasonable. Then, the basic task segments are written into the task storage area, which stores tasks that require priority processing and have clear execution rules; the flexible task segments are written into the extended task buffer, awaiting further scheduling and optimization based on actual conditions, so as to better adapt to various changes in the production process and improve production efficiency and resource utilization.

[0075] Example 3:

[0076] In the cluster status analysis module, different types of preset constraint matrices will result in different scheduling level generation methods. When the preset constraint matrix adopts a static analysis model, the scheduling level is the interval mapping result of the linear superposition value of the first task topology model and the second task topology model. In this case, the first task topology model and the second task topology model are first linearly superimposed. Linear superposition is a simple and direct fusion method, which adds the relevant data of the two models to obtain a comprehensive data value. Then, this linear superposition value is interval mapped. Interval mapping is to map the linear superposition value to the corresponding scheduling level interval based on a pre-set interval range, thereby determining the current scheduling level. This method is suitable for scenarios where tasks and resources are relatively stable, and can quickly generate scheduling levels.

[0077] When the preset constraint matrix adopts a dynamic adjustment model, the scheduling level is a continuous set of states that uses an incremental learning algorithm to modify the associated data between the first and second task topology models in real time. This incremental learning algorithm gradually updates and optimizes the model as new data is continuously input. During this process, the associated data between the first and second task topology models is continuously analyzed and learned, and the scheduling level is modified in real time based on new circumstances. This approach can better adapt to the dynamic changes in tasks and resources, improving the accuracy and rationality of task allocation.

[0078] Consider a large intelligent logistics park with numerous cargo handling and storage tasks, performed by an AMR cluster. Within this logistics park, cargo handling and storage tasks continuously generate new demand data. This data is processed by the task scheduling module and the twin modeling module to generate the first task topology model M1 and the second task topology model M2.

[0079] When the static analysis model is used for the preset constraint matrix, a logistics park receives a batch of e-commerce promotional goods handling and storage tasks within a certain time period. The first task topology model, M1, primarily reflects the distribution and resource requirements of basic handling tasks in three-dimensional space, such as the relationship between each cargo storage point and the target storage point, and the number of AMRs required for different handling paths. The second task topology model, M2, reflects dynamic task adjustments and resource matching, such as the impact of temporary expedited orders on handling task priorities and resource allocation.

[0080] First, linearly superimpose M1 and M2. The formula for linear superposition is S = M1 + M2, where S represents the value after linear superposition, and M1 and M2 represent the first task topology model and the second task topology model, respectively. This formula simply adds the various data information contained in the two models. For example, the resource demand value corresponding to each task node in M1 is added to the resource demand value of the corresponding node in M2 to obtain a new comprehensive resource demand value. The association weight values between the tasks in the two models are also added to combine the characteristics of the two models.

[0081] After obtaining the linear superposition value S, interval mapping is performed. Assume that based on the logistics park's long-term operational data and experience, three scheduling level intervals have been established: low-priority tasks are assigned a range of [0, 100), medium-priority tasks are assigned a range of [100, 300), and high-priority tasks are assigned a range of [300, +∞). The value of S is mapped against these intervals. If S = 150, then according to the interval mapping rules, the scheduling level for this batch of tasks is determined to be medium priority. This allows the logistics park to prioritize resources for medium-priority tasks based on this scheduling level, such as allocating an appropriate number of AMR robots to handle the handling and storage of goods for this e-commerce promotion, ensuring the orderly progress of tasks.

[0082] When the preset constraint matrix uses a dynamic adjustment model: During daily logistics park operations, order status fluctuates constantly, and warehouse layouts may also change due to changes in cargo types. For example, on a given weekday, a warehouse area originally reserved for daily necessities may need to temporarily store a large amount of fresh produce. This results in changes to warehouse storage tasks and transportation routes, and the corresponding task topology model will also change.

[0083] At this point, the cloud platform uses an incremental learning algorithm to make real-time corrections to the associated data between M1 and M2. The incremental learning algorithm continuously receives new task data and equipment status information, such as the weight and volume of new orders, and the battery level and location of the AMR robots. For example, suppose a new batch of heavy industrial equipment handling tasks arrives at a certain moment. This increases the demand for large AMR robots and may alter the handling routes. Based on this new information, the incremental learning algorithm adjusts the relevant task nodes, task relationships, and resource requirements in M1 and M2.

[0084] During this dynamic adjustment process, the scheduling level is no longer a fixed interval mapping result, but a continuous set of states that is continuously updated through an incremental learning algorithm. Each input of new data causes the scheduling level to change accordingly to adapt to the ever-changing tasks and resource conditions within the logistics park. For example, as new tasks are continuously added and equipment status is updated in real time, the scheduling level may gradually evolve from an initial state to a new state that better reflects the current situation. This new state comprehensively considers all the latest task and resource information, thereby achieving precise allocation and efficient scheduling of AMR cluster tasks, ensuring the efficient operation of the logistics park.

[0085] Example 4:

[0086] The cloud platform also includes a device collaboration module connected to the core computing unit and connected to the physical device resource pool via a data exchange interface. Based on the device scheduling requirements in the final scheduling instructions, the device collaboration module selects available device queues from the physical device resource pool and generates task execution sequences to optimize cluster operation processes.

[0087] When generating a task execution sequence, the 3D spatial topology model is first loaded. This 3D spatial topology model can intuitively display the spatial layout and positional relationships of physical devices. The spatial coordinates of each device in the available device queue are located in the topology model, accurately obtaining the device's current location information. Then, the optimal movement trajectory from each device's current location to the target work area is calculated based on a dynamic programming algorithm. The dynamic programming algorithm is an efficient optimization algorithm that can find the optimal path in complex spatial environments. At the same time, the available device queue is sorted by task urgency to ensure that urgent tasks are processed first. Finally, the optimal movement trajectory and efficiency ranking are integrated into the topology model to generate a visual task execution sequence. This visual task execution sequence allows operators to more intuitively understand the task execution process and equipment operating status, facilitating monitoring and management.

[0088] Take a large electronics manufacturing plant as an example. The plant has multiple production lines, and every day a large number of tasks such as raw material handling, parts assembly, and finished product transportation need to be completed by the AMR cluster. This highly relies on the AMR cluster task planning cloud platform based on digital twins for efficient scheduling.

[0089] After receiving the task requirement data, the factory's cloud platform generates the final scheduling instructions through a series of processing. This is when the device collaboration module, connected to the core computing unit, comes into play. In this factory, the physical device resource pool includes AMRs of different types and functions. For example, some AMRs are responsible for moving materials between the raw material warehouse and the production line, while others are specifically used to transfer parts between different workstations on the production line. Based on the device scheduling requirements in the final scheduling instructions, the device collaboration module selects a queue of available devices from the physical device resource pool.

[0090] Suppose there's an urgent production task requiring the quick transport of a specific batch of raw materials from the warehouse to a designated workstation on the production line. The device coordination module first screens the AMRs in the physical device resource pool, taking into account factors such as the AMR's current location, power level, and load capacity. For example, AMRs that are close to the raw material warehouse, have sufficient power, and sufficient load capacity for the task are prioritized and formed into the available device queue.

[0091] The device collaboration module generates task execution sequences to optimize cluster workflows. It loads a three-dimensional spatial topology model that details the layout of each area within the factory, including the raw material warehouse, production lines, temporary storage areas, and the locations of individual devices. Using this model, the device collaboration module can precisely locate the spatial coordinates of each AMR in the available device queue, providing a clear understanding of each AMR's current position.

[0092] A dynamic programming algorithm calculates the optimal trajectory from each AMR's current position to the target work area (i.e., the production line workstation to which the raw materials need to be moved). The dynamic programming algorithm comprehensively considers the factory layout, such as avoiding obstacles and selecting the shortest or smoothest path. For example, if a passage in a certain area of the factory is temporarily inaccessible due to equipment maintenance, the dynamic programming algorithm will automatically avoid this area and plan an alternative suitable route for the AMR.

[0093] The equipment coordination module prioritizes the available equipment by task urgency. For urgent production tasks like those mentioned above, AMRs that can complete handling tasks more quickly are prioritized. For example, AMRs with higher speed and efficiency are prioritized to ensure that raw materials reach the production line as quickly as possible without delaying production.

[0094] The equipment collaboration module integrates the calculated optimal movement trajectory and efficiency ranking into a three-dimensional spatial topology model to generate a visual task execution sequence. This visual task execution sequence intuitively displays the movement route and execution order of each AMR, allowing factory managers to monitor task progress in real time. Through this visual interface, managers can clearly see which AMR is heading to the raw material warehouse, which AMR has loaded raw materials and is transporting them to the production line, and the estimated completion time of the entire handling task. If an abnormality occurs during task execution, such as an AMR malfunction, managers can promptly detect it and readjust task allocation based on the visual information to ensure the smooth progress of production tasks.

[0095] Example 5:

[0096] The core computing unit uses a dual verification model, combining an abnormality detection mechanism and a constraint conflict verification mechanism, to perform logical checks on task assignment benchmarks and navigation parameters. The abnormality detection mechanism verifies the operational integrity of equipment. During task execution, equipment may experience various failures or abnormalities. The abnormality detection mechanism monitors and analyzes equipment operating data in real time to determine whether the equipment is operating normally. If an equipment anomaly is detected, appropriate measures are taken promptly, such as adjusting task assignments or performing equipment repairs, to ensure smooth mission execution.

[0097] The constraint conflict check mechanism is used to resolve temporal and spatial conflicts between tasks. In multi-tasking scenarios, temporal and spatial conflicts may arise between tasks. For example, two tasks may require the same device at the same time, or their execution paths may conflict. The constraint conflict check mechanism analyzes and coordinates the timing and spatial requirements of tasks to avoid conflicts and ensure orderly execution.

[0098] The cloud platform also includes an instruction conversion module and a task archiving module, connected to the core computing unit. The instruction conversion module converts the final scheduling instructions into device control codes and sends them to the designated device cluster via a data exchange interface to initiate the task program. The task archiving module stores job requirement data, the first and second task topology models, the scheduling level, and the final scheduling instructions, and generates a complete job record chain based on the task lifecycle. These modules work together to ensure the efficient operation of the cloud platform and the smooth execution of tasks.

[0099] For example, at a large express delivery sorting center, tens of thousands of packages are sorted and delivered daily. Within this center, a cluster of AMRs handles tasks such as package handling, sorting, and loading. After generating the final dispatch instructions, the core computing unit performs a logical check on the task allocation benchmark and navigation parameters, employing a dual verification mechanism: abnormal state detection and constraint conflict verification.

[0100] The abnormal state detection mechanism is used to verify the operational integrity of equipment. In express sorting centers, AMRs (Automated Automated Response Machines) are required to operate continuously for extended periods of time, and equipment failures can occur at any time. For example, while an AMR is handling a package, its battery monitoring system sends data to the cloud platform. The abnormal state detection mechanism analyzes this data in real time. If it detects an abnormal drop in battery power—for example, if the battery consumption for a single handling task normally remains below 10% and the AMR's battery drops by more than 15%—the system immediately identifies the device as abnormal. The abnormal state detection mechanism also monitors AMR sensor data, such as LiDAR and visual sensors. If LiDAR data indicates unusual fluctuations in its scanning range, this could indicate sensor interference or failure, triggering the abnormal state detection mechanism. Once an anomaly is detected, the cloud platform immediately issues an alert, notifies technicians for repairs, and reassigns the AMR's tasks to other, functioning devices, ensuring minimal disruption to package sorting operations.

[0101] The constraint conflict check mechanism is used to resolve temporal and spatial conflicts between tasks. In express delivery sorting centers, numerous AMRs (Automated Automated Machines) operate simultaneously within a limited space. Improper task planning can easily lead to temporal and spatial conflicts. For example, at a given moment, two AMRs receive separate tasks: one to move a batch of packages from sorting area A to loading area X, and the other to move another batch of packages from sorting area B to loading area Y. Their planned paths intersect in a narrow aisle. The constraint conflict check mechanism analyzes the task paths and execution times of all AMRs before executing a task. When a conflict is detected, the system adjusts the task path based on factors such as the urgency of the package and the AMR's current location. If the task of moving the package from sorting area A is more urgent, the system adjusts the path of the other AMR, forcing it to take a slight detour to avoid the conflicting area. This ensures smooth execution of both tasks and prevents AMRs from waiting or colliding in the aisle.

[0102] The cloud platform also includes an instruction conversion module and a task archiving module. The instruction conversion module converts the final dispatch instruction into a device control code and sends it to the designated device cluster via a data exchange interface to initiate the task. In an express sorting center, the final dispatch instruction might be "Have AMR number 001 move 50 packages from sorting area C to loading area Z." The instruction conversion module converts this instruction into a control code recognizable by the AMR, such as a series of numbers and letters, and sends it to AMR number 001 via the data exchange interface. After receiving the control code, the AMR initiates the transport task according to the instruction, accurately driving to sorting area C, picking up 50 packages, and then transporting them to loading area Z.

[0103] The task archiving module stores job requirement data, the primary and secondary task topology models, the scheduling level, and the final dispatch instructions. It also generates a complete job record chain based on the task lifecycle. In the daily operations of an express sorting center, detailed records are kept for every package handling task. For example, for a task involving moving packages from sorting area D to loading area W, the task archiving module records the job requirement data, including the number of packages, weight, and destination; the task-related components of the primary and secondary task topology models, such as the task's layout in three-dimensional space and resource matching; the scheduling level, or the task's priority; and the final dispatch instructions. This information is organized along the task lifecycle, from task generation, assignment, execution, to completion, forming a complete job record chain. By reviewing these records, managers can trace the execution of tasks, analyze the rationality of task planning, and identify lessons learned to optimize subsequent task scheduling and management strategies, ultimately improving the overall operational efficiency of the express sorting center.

[0104] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0105] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. AMR cluster task planning cloud platform based on digital twin, characterized by: include: A task scheduling module is used to receive multi-source job demand data and identify task types, and divide the job demand data into a basic task data stream and an extended task data stream based on preset task decomposition rules; a twin modeling module, configured to perform three-dimensional spatial gridding processing on the basic task data stream to generate a first task topology model, and perform dynamic resource matching processing on the extended task data stream to generate a second task topology model; A cluster status analysis module is used to fuse the first task topology model and the second task topology model according to a preset constraint matrix, and then map them to corresponding scheduling levels, and use the scheduling levels as the task allocation benchmark for the current cluster; A path planning module, configured to call an optimal path solution from a preset path database based on the scheduling level, and use the optimal path solution as a navigation parameter for the current cluster; The core computing unit is used to perform logic verification on the task allocation benchmark and the navigation parameter to generate a final scheduling instruction.

2. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: The twin modeling module performs dynamic resource matching processing on the extended task data stream, including: Dividing the task sequence in the extended task data stream into a fixed task group and a flexible task group, and calculating the resource occupancy rate of the flexible task group based on a preset dynamic load balancing algorithm to generate a load feature set; Performing hierarchical processing on the device status data in the extended task data stream, extracting the energy consumption characteristics of each layer of equipment and constructing an energy consumption map; Dynamically associate and match the load feature set with the energy consumption spectrum to generate the second task topology model.

3. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: The preset task decomposition rules include a rigid constraint set and a flexible constraint set; the rigid constraint set includes a time window identifier, an equipment capacity identifier, and a job priority identifier; the flexible constraint set includes an energy consumption threshold identifier and a path conflict identifier, and each identifier corresponds to an independent task processing channel.

4. The AMR cluster task planning cloud platform based on digital twin according to claim 3 is characterized in that: The platform also includes a data interaction interface, which is used to realize communication docking between the task scheduling module, the twin modeling module, the cluster state analysis module and the path planning module and the physical device network respectively; The task scheduling module divides the job demand data based on the preset task decomposition rules, including: Acquire a mixed data packet from a physical device network in real time through the data interaction interface, and match the protocol label of the mixed data packet according to the identifier in the rigid constraint set to separate the basic task segments; Traversing the extended tags of the mixed data packet according to the identifier in the flexible constraint set to extract the flexible task segment; The basic task segment and the flexible task segment are synchronized and aligned according to the task triggering time and then written into the task storage area and the extended task buffer area respectively.

5. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: When the preset constraint matrix adopts a static analysis model, the scheduling level is an interval mapping result of the linear superposition value of the first task topology model and the second task topology model; When the preset constraint matrix adopts a dynamic adjustment model, the scheduling level is a continuous state set in which the associated data of the first task topology model and the second task topology model are corrected in real time through an incremental learning algorithm.

6. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: It also includes a device collaboration module connected to the core computing unit, and the device collaboration module is connected to the physical device resource pool through the data interaction interface; The device coordination module is configured to screen available device queues from the physical device resource pool according to the device scheduling requirements in the final scheduling instruction, and generate a task execution sequence to optimize the cluster operation process.

7. The AMR cluster task planning cloud platform based on digital twin according to claim 6 is characterized in that: The device collaboration module generates a task execution sequence including: Loading a three-dimensional spatial topology model, and locating the spatial coordinate point of each device in the available device queue in the topology model; Calculate the optimal movement trajectory from the current position of each device to the target operation area based on a dynamic programming algorithm, and sort the available device queues by their urgency; The optimal movement trajectory and the efficiency ranking are integrated into the topology model to generate a visualization task execution sequence.

8. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: When the core computing unit performs logical verification on the task allocation benchmark and navigation parameters, it adopts a dual verification mode of abnormal state detection mechanism and constraint conflict verification mechanism. The abnormal state detection mechanism is used to confirm the integrity of equipment operation, and the constraint conflict verification mechanism is used to resolve the temporal and spatial contradictions between tasks.

9. The AMR cluster task planning cloud platform based on digital twin according to claim 1 is characterized in that: It also includes an instruction conversion module connected to the core computing unit, which is used to convert the final scheduling instruction into a device control code and send the device control code to the designated device cluster through the data interaction interface to start the task program.

10. The AMR cluster task planning cloud platform based on digital twin according to claim 1, characterized in that: It also includes a task archiving module connected to the core computing unit, which is used to store the job demand data, the first task topology model, the second task topology model, the scheduling level and the final scheduling instruction, and generate a complete job record chain according to the task life cycle.

Images