A multi-aelos robot collaboration sorting method and related device for warehouse logistics

Abstract

The application provides a multi-Aelos robot cooperative sorting method and related device for warehouse logistics, and relates to the technical field of warehouse logistics. Real-time order data is received, and the real-time order data is analyzed to generate a task set containing cargo location, target sorting area and priority; through a distributed task allocation mechanism, the task matching degree is dynamically calculated in combination with the state information of each robot itself, and a reinforcement learning model is used to optimize the global task allocation to generate a global task allocation result; on the basis of the global allocation strategy, the warehouse environment is modeled as a space-time graph structure; a pathfinding algorithm is used in combination with the space-time graph structure to generate a conflict-free path and optimize the trajectory in combination with the task time limit constraint; according to the target position in the conflict-free path, the cargo is identified through a vision system, and the grabbing and placing are completed in combination with a force control mechanical arm. The technical effects of improving sorting efficiency and adaptability and reducing operation and maintenance cost are realized.

Description

Technical Field

[0001] This application relates to the field of warehousing and logistics technology, and in particular to a multi-Aelos robot collaborative sorting method and related apparatus for warehousing and logistics. Background Technology

[0002] In the warehousing and logistics sector, goods sorting is a core component, and its efficiency directly determines the overall operating costs and customer satisfaction of the logistics system. Currently, traditional sorting solutions mainly rely on manual operation or fixed automated equipment, but these have significant limitations in practical applications. While manual sorting offers flexibility, it is inefficient and has a high error rate, especially when orders surge, highlighting its bottlenecks. With rising labor costs, logistics systems reliant on manual labor face enormous economic pressure. While fixed automated equipment (such as conveyor belts and robotic arms) can replace some human labor, their rigid design results in insufficient flexibility and adaptability. These devices are typically developed for specific tasks and struggle to cope with dynamic warehousing environments and diverse order demands. High deployment and maintenance costs, along with stringent spatial requirements, further increase the difficulty of modifications during business expansion or order fluctuations.

[0003] In recent years, multi-robot collaborative systems have become a research hotspot, significantly improving sorting efficiency and flexibility through distributed task allocation and collaborative operations. However, existing technologies still face many bottlenecks: centralized task allocation relies on a single control node, resulting in high communication overhead and poor real-time performance, making it difficult to adapt to dynamic task changes; traditional path planning algorithms (such as A / B) also face challenges. Dijkstra's approach ignores the time dimension, leading to frequent multi-machine path conflicts and collisions; collaborative control relies on preset rules, making it difficult to meet the high-requirement scenarios in terms of handling accuracy and stability. Although some solutions introduce automation technology, they still generally suffer from drawbacks such as high risk of single-point failure, frequent path conflicts, and insufficient adaptive capabilities.

[0004] With the rapid development of artificial intelligence, the Internet of Things, and robotics, warehousing and logistics are accelerating their transformation towards intelligence and flexibility. The market's demand for efficient and low-cost dynamic sorting is becoming increasingly urgent, requiring an innovative solution that integrates dynamic task allocation, spatiotemporal path planning, and precise collaborative control. Summary of the Invention

[0005] To meet the needs of efficient and low-cost dynamic sorting, this application provides a multi-Aelos robot collaborative sorting method and related apparatus for warehouse logistics.

[0006] Firstly, this application provides a multi-Aelos robot collaborative sorting method for warehousing and logistics, which adopts the following technical solution: A collaborative sorting method using multiple Aelos robots for warehouse logistics includes: Receive real-time order data and parse the real-time order data to generate a task set that includes the location of goods, target sorting area and priority. Through a distributed task allocation mechanism, the task matching degree is dynamically calculated by combining the state information of each robot, and a reinforcement learning model is used to optimize the global task allocation and generate the global task allocation result. Based on the global allocation strategy, the warehouse environment is modeled as a spatiotemporal graph structure; A pathfinding algorithm is used in conjunction with the aforementioned spatiotemporal graph structure to generate conflict-free paths, and the trajectory is optimized in conjunction with task time constraints. Based on the target location in the conflict-free path, the goods are identified by a vision system and then grasped and placed by a force-controlled robotic arm.

[0007] Optionally, the generated task set includes: Automatically identifies the weight, dimensions, and fragility characteristics of goods; Promotional orders are automatically marked as the highest priority and assigned to an emergency processing channel; The task queue is refreshed immediately when a new order arrives to ensure timeliness.

[0008] Optionally, dynamically calculating the task matching degree includes: A multi-dimensional status assessment is performed by integrating the robot's real-time location, remaining battery power, current load, and task urgency. Automatically adjust the weighting of the allocation strategy based on historical sorting data to achieve reinforcement learning optimization; Set up fast lanes for high-priority orders to skip the regular queuing process and enable competitive task allocation.

[0009] Optionally, constructing the spatiotemporal graph structure includes: Add timestamps to key locations for dynamic node marking; Automatically increase passage costs based on real-time robot density in the passageway; Reserve dedicated passage periods for emergency tasks to achieve time window constraint integration.

[0010] Optionally, generating conflict-free paths includes: A hierarchical planning strategy is adopted, using spatiotemporal A at the global level. The algorithm generates the main path and avoids dynamic obstacles in real time at the local layer using a potential field model. Insert smooth curve segments into sharp turns to ensure stable robot steering and achieve smooth trajectory processing; Assign high-value order robots a direct fast lane to achieve priority passage.

[0011] Optionally, the grab and place operations include: Vision-force control dual closed loop: The vision system locates the coordinates of the goods at the millimeter level, and the force-controlled robotic arm automatically adjusts the gripping force according to the characteristics of the goods; Multi-robot collaborative collision avoidance: The central controller allocates staggered grasping time slots for adjacent robots; Special handling for fragile items: Automatically activates a buffer gripping mode for glass products, etc.

[0012] Optionally, the method further includes intelligent fault response: Multi-level early warning mechanism: A level three alarm is triggered when the battery level is below 20% or the temperature exceeds 70°C; Seamless task transfer: Automatically transfer tasks based on the location, idle status, and load capacity of nearby robots; Load balancing guarantee: Heavy tasks are intelligently split into multiple sub-tasks for distribution.

[0013] Secondly, this application provides a multi-Aelos robot collaborative sorting device for warehousing and logistics, comprising: The data acquisition module is used to receive real-time order data and parse the real-time order data to generate a task set containing the location of goods, target sorting area and priority. The result allocation module is used to dynamically calculate the task matching degree by combining the state information of each robot with the distributed task allocation mechanism, and to optimize the global task allocation by using a reinforcement learning model to generate the global task allocation result. The spatiotemporal graph module is used to model the warehouse environment as a spatiotemporal graph structure based on a global allocation strategy; The optimization module is used to generate conflict-free paths by combining the pathfinding algorithm with the spatiotemporal graph structure and optimize the trajectory by combining the task time constraints. The execution module is used to identify the goods through a vision system based on the target position in the conflict-free path, and then combine it with a force-controlled robotic arm to complete the grasping and placement.

[0014] Thirdly, this application provides a computer device, the device comprising: a memory and a processor, wherein the processor, when executing computer instructions stored in the memory, performs the method described above.

[0015] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the method described above.

[0016] In summary, this application receives real-time order data, parses the data to generate a task set containing the location of goods, target sorting area, and priority; uses a distributed task allocation mechanism, dynamically calculates task matching degree by combining the state information of each robot, and employs a reinforcement learning model to optimize global task allocation and generate a global task allocation result; based on the global allocation strategy, models the warehouse environment as a spatiotemporal graph structure; uses a pathfinding algorithm combined with the spatiotemporal graph structure to generate conflict-free paths and optimizes the trajectory by combining task time constraints; based on the target position in the conflict-free path, identifies goods through a vision system, and completes grasping and placement using a force-controlled robotic arm. This achieves the technical effects of improving sorting efficiency and adaptability while reducing operation and maintenance costs. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application; Figure 2 This is a flowchart illustrating the first embodiment of the multi-Aelos robot collaborative sorting method for warehousing and logistics in this application; Figure 3 This is a flowchart of the dynamic task allocation algorithm in this application; Figure 4 This is a flowchart of the spatiotemporal path planning algorithm in this application; Figure 5 This is a structural block diagram of the first embodiment of the multi-Aelos robot collaborative sorting device for warehousing and logistics in this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] Reference Figure 1 , Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application.

[0020] like Figure 1As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0021] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0022] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a multi-Aelos robot collaborative sorting program for warehousing and logistics.

[0023] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in this application can be set in the computer device. The computer device calls the multi-Aelos robot collaborative sorting program for warehousing and logistics stored in the memory 1005 through the processor 1001, and executes the multi-Aelos robot collaborative sorting method for warehousing and logistics provided in the embodiment of this application.

[0024] This application provides a multi-Aelos robot collaborative sorting method for warehousing and logistics, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the multi-Aelos robot collaborative sorting method for warehousing and logistics in this application.

[0025] In this embodiment, the multi-Aelos robot collaborative sorting method for warehousing and logistics includes the following steps: Step S10: Receive real-time order data and parse the real-time order data to generate a task set containing the location of goods, target sorting area and priority.

[0026] It should be noted that the flowchart of the dynamic task allocation algorithm is as follows: Figure 3 As shown, the generated task set includes: automatically identifying the weight, size, and fragility characteristics of goods; automatically marking promotional orders as the highest priority and assigning them to an emergency processing channel; and immediately refreshing the task queue when new orders arrive to ensure timeliness.

[0027] In practice, the central control system receives real-time order data from the warehouse management system, parses it to generate a set of sorting tasks, each task containing key parameters such as goods location, target sorting area, priority, and goods attributes. The system employs a distributed task allocation mechanism.

[0028] Step S20: Through a distributed task allocation mechanism, the task matching degree is dynamically calculated by combining the state information of each robot, and a reinforcement learning model is used to optimize the global task allocation and generate the global task allocation result.

[0029] In practice, the dynamic calculation of task matching degree includes: multi-dimensional state evaluation based on the robot's real-time position, remaining power, current load and task urgency; automatic adjustment of allocation strategy weights based on historical sorting data to achieve reinforcement learning optimization; and setting up fast channels for high-priority orders to skip the regular queuing process to achieve bidding-style task allocation.

[0030] It's important to note that each Aelos robot dynamically calculates its task matching degree based on its real-time status (such as location, remaining battery power, and load capacity) and submits the task value assessment results to the central system through a bidding strategy. The central system optimizes global task allocation decisions based on a reinforcement learning model, prioritizing the allocation of high-priority tasks to the most efficient robots. Simultaneously, it triggers real-time task reassignment in case of order changes or robot malfunctions, ensuring the continuity and responsiveness of the sorting process. For example, when a delay in a promotional order is detected, the system automatically adjusts the allocation weights, strengthens time constraints, and ensures that urgent tasks are completed first.

[0031] It should be noted that task modeling and state representation include: each sorting task is defined as a quadruple containing spatiotemporal constraints: The location of the goods Real-time positioning via laser SLAM (accuracy ±2cm), task time limit Dynamically set according to the urgency of the order (e.g., expedited orders) (seconds), priority Distinguish between task importance (value range 1-5) and cargo attributes. Record weight, dimensions, and special handling requirements. Robot state is represented by real-time vector data. in Updated by UWB indoor positioning system (frequency 50Hz). Monitor remaining battery power. Record the current load (maximum load) ), This forms a queue of tasks to be executed. These parameters are synchronized to the central controller at a frequency of 10Hz, providing a real-time data basis for dynamic decision-making.

[0032] Step S30: Based on the global allocation strategy, model the warehouse environment as a spatiotemporal graph structure.

[0033] It should be noted that constructing the spatiotemporal graph structure includes: adding timestamps to key locations for dynamic node marking; automatically increasing passage costs based on real-time robot density in the passage; and reserving dedicated passage periods for emergency tasks to achieve time window constraint integration.

[0034] It should be noted that during the sorting process, the system models the warehouse environment as a spatiotemporal graph structure. Nodes at key locations (such as shelves and sorting areas) are appended with timestamps to record robot occupancy status, and edge weights dynamically reflect path congestion levels. Global path planning employs an improved spatiotemporal A... The algorithm, combining time window constraints and curvature smoothing optimization, generates conflict-free paths. The robot uses LiDAR and visual sensors to perceive dynamic obstacles (such as other robots or temporarily stacked goods) in real time, adjusting its trajectory using a potential field model to avoid collisions and optimize passage efficiency. For example, when a passageway is congested, the system automatically guides the robot to detour or wait, ensuring high-priority tasks pass first, ultimately reducing the path conflict rate from 12% in traditional systems to 3.5%.

[0035] In practical implementation, the core of task allocation is the computational robot. With the task Match degree: This function optimizes spatial distance, time urgency, and load balancing through a three-factor weighted sum. Weighting coefficients The initial value is dynamically adjusted through reinforcement learning, and its reward function is designed as follows: The first reward is the timely completion of high-priority tasks (delay time). The exponential decay of the algorithm strengthens the time constraint, the second term penalizes load imbalance, and the last term suppresses path conflict. The Q-learning algorithm uses... -greedy strategy ( After 1000 iterations of training, the system converges to a stable policy, enabling it to adaptively adjust weight parameters. For example, when a high-priority task's latency rate exceeds a threshold, it automatically increases the priority level. The weight is increased to 0.4 to strengthen the time constraint.

[0036] It should be noted that the actual task allocation adopts a bidding mechanism with time and space constraints as follows: Among them, waiting time The exponential decay term prevents task starvation, and the emergency task bias term ( (For the indicator function) to ensure priority response to critical orders. The central controller runs the Hungarian algorithm every 200ms to solve for the optimal allocation matrix, while simultaneously monitoring multiple condition-triggered events: When the robot's battery level drops below 20%, the remaining task time is less than 60 seconds, or there are more than 3 collisions in a single area, the system will initiate task reallocation within 50ms and update the environmental access costs. This formula uses the congestion coefficient ( ) and channel capacity limitations ( Dynamically adjust path weights. For example, when three robots are simultaneously present in a certain passage, the passage cost increases to [a certain value]. This guides subsequent robots to choose alternative paths.

[0037] Step S40: Use a pathfinding algorithm combined with the spatiotemporal graph structure to generate a conflict-free path and optimize the trajectory in combination with task time constraints.

[0038] In practical implementation, the flowchart of the spatiotemporal path planning algorithm is as follows: Figure 4 As shown. Generating conflict-free paths includes: a hierarchical planning strategy, employing spatiotemporal A at the global level. The algorithm generates the main path and avoids dynamic obstacles in real time at the local layer using a potential field model; it inserts smooth curve segments into sharp turns to ensure stable robot turning and achieve trajectory smoothing; and it allocates straight fast lanes to robots handling high-value orders to achieve priority passage.

[0039] In practical implementation, spatiotemporal path planning and dynamic obstacle avoidance include: First, the warehousing environment is abstracted into a spatiotemporal diagram structure. , where nodes Indicates key locations such as shelves, sorting areas, and charging stations, and includes a timestamp. To record the robot's occupancy status (such as the time windows when the robot enters and leaves the node). Edge Connecting adjacent nodes, their weights The real-time travel cost of the route is dynamically reflected, and the calculation formula is as follows: in, The basic path length. Indicates time window The number of robots that pass through this path. This is the channel capacity threshold.

[0040] Global path planning employs an improved spatiotemporal A The algorithm ensures conflict-free multi-robot paths by introducing time window constraints and curvature smoothing optimization. The heuristic function is designed as follows: The first term represents the Euclidean distance from the current node to the target, and the second term is a time penalty term used to strengthen the time constraints of urgent tasks. If the path curvature... Exceeding the threshold Smooth steering is achieved by inserting B-spline curve control points, ensuring robot motion stability. The curvature calculation formula after path smoothing is as follows: Local dynamic obstacle avoidance combines real-time perception with a potential field model to resolve interference from temporary obstacles (such as other robots or moving goods). The robot acquires environmental point cloud data through LiDAR and visual sensors, and constructs a local obstacle map using an occupancy grid mapping method, with grid occupancy probability... An obstacle is marked if the force exceeds 0.7. The trajectory adjustment is based on a virtual potential field model, with the total force... It consists of attractive force, repulsive force, and traffic guiding force: Attraction: Guides the robot to move towards the target along a global path; the calculation formula is: Repulsive force: used to avoid obstacles; the calculation formula is: Traffic guidance force: Dynamically adjusts path weights based on the degree of congestion in the passageway to guide the robot to detour around high-density areas. The calculation formula is as follows: To enhance dynamic adaptability, the system updates the environmental state every 200ms and shares the real-time position and task progress of each robot via the ROS communication network. When a path conflict or temporary obstacle is detected, the central controller triggers local path replanning within 50ms, generating a smooth detour trajectory based on the potential field model. For example, in a congested passage scenario, the system automatically guides low-priority task robots into the waiting queue while allocating fast passage paths for high-priority task robots, reducing the overall path conflict rate from 12% in traditional systems to 3.5%.

[0041] Furthermore, the path planning module is deeply integrated with task allocation and fault recovery mechanisms. When a robot's task is interrupted due to a malfunction, the system not only reassigns the task but also adjusts the path weights of other robots in real time based on the updated spatiotemporal graph to avoid resource conflicts. Simulation tests show that when 50 robots work collaboratively, the dynamic obstacle avoidance response time is less than 0.5 seconds, and the path planning success rate exceeds 98%, significantly improving the robustness and efficiency of the sorting system.

[0042] Step S50: Based on the target position in the conflict-free path, identify the goods through the vision system, and complete the grasping and placement in conjunction with the force-controlled robotic arm.

[0043] It should be noted that the grasping and placement operations include: vision-force control dual closed loop: the vision system locates the coordinates of the goods at the millimeter level, and the force control robotic arm automatically adjusts the gripping force according to the characteristics of the goods; multi-machine collaborative anti-collision: the central controller allocates staggered grasping time slots for adjacent robots; special handling of fragile items: the buffer grasping mode is automatically activated for glass products and the like.

[0044] In practice, the robot identifies target goods using a high-precision vision system, combines this with deep learning algorithms to match the goods' feature database, and locates the precise coordinates on the shelf. A six-axis force-controlled robotic arm dynamically adjusts its gripping force, achieving stable grasping and placement of goods through multi-sensor fusion technology (visual feedback, torque detection, and inertial navigation). After being transported to the sorting area, the robot categorizes the goods according to order information and places them in designated outbound areas, ensuring millimeter-level sorting accuracy through a visual guidance and positioning system. For example, when grasping fragile items, the system automatically reduces the gripping force to prevent damage.

[0045] It should be noted that the multi-robot communication network built on the ROS framework shares the status information of each robot in real time (such as position, task progress, and remaining battery power). When a robot malfunctions due to insufficient battery power, excessive temperature, or communication delay, the central system automatically activates the fault response mechanism. Based on the remaining load capacity and task priority of nearby robots, the system selects the optimal support robot to take over the task. For example, when the faulty robot has a load of 50kg, the system prioritizes allocating 30kg to a nearby idle robot, with the remaining 20kg shared by other robots. At the same time, the global path planning is updated to avoid resource conflicts, and the fault recovery time is controlled within 8.5 seconds.

[0046] It is understood that the method also includes intelligent fault response: multi-level early warning mechanism: triggering a level 3 alarm when the battery level is below 20% or the temperature exceeds 70°C; seamless task transfer: automatically transferring tasks based on the location of nearby robots, their idle status, and their load capacity; load balancing guarantee: intelligently splitting overloaded tasks into multiple sub-tasks for distribution.

[0047] In practical implementation, visual localization and 3D coordinate mapping include: the robot uses a binocular vision system to capture RGB-D images of the shelf area in real time, and combines this with an improved ORB-SLAM algorithm to construct a high-precision environmental point cloud map. The pixel coordinates of the target goods are also considered. Through camera intrinsic parameter matrix Mapping to three-dimensional space coordinates : in, For camera focal length, The system uses the master point coordinates. For goods with high reflectivity or low texture, a multi-frequency stripe structured light projection technology is introduced to compensate for the lack of depth information through phase calculation. Combined with the Harris-Laplace corner detection algorithm, sub-pixel-level edge features are extracted, keeping the positioning error within a certain range. Within this range. In dynamic warehousing environments, ORB-SLAM updates cargo pose in real time through keyframe matching and local map optimization, eliminating positioning drift caused by cargo displacement or stacking deformation.

[0048] Adaptive force control algorithms and clamping optimizations include: The six-axis force-controlled robotic arm adjusts the gripping force based on a dynamic PID model, targeting the desired gripping force. The settings are dynamically adjusted based on the weight, material, and fragility of the goods. The adaptive PID controller is designed as follows: Among them, the error term is: Real-time feedback is provided by a six-dimensional torque sensor, and high-frequency noise is suppressed by a Kalman filter. Gripper opening and closing angle. The mapping relationship with the target clamping force is determined by a pre-calibrated nonlinear function: Achieve, ensuring clamping force is to Smooth transition within the range. The system automatically switches control modes according to different cargo characteristics: increasing the proportional coefficient when gripping rigid cargo. To enhance anti-interference capabilities; to strengthen the points component when handling fragile goods. To eliminate steady-state error.

[0049] Multimodal perception fusion and dynamic correction include: Stability during the grasping process depends on the deep integration of visual, torque feedback, and inertial navigation data.

[0050] Visual servo correction: Real-time tracking of cargo pose offset using binocular vision. Generate joint space compensation instructions: in, Let be the Jacobian matrix of the robotic arm. A Lie group spatial interpolation algorithm is used to ensure a smooth and continuous compensation trajectory, avoiding jitter caused by the kinematic singularities of the robotic arm.

[0051] Inertial anti-shake control: based on IMU detection of vibration acceleration at the end of the robotic arm Frequency domain notch filter is used to suppress The above refers to high-frequency jitter.

[0052] Contact force sensing: When the torque sensor detects a sudden change in contact force (such as cargo slippage), an emergency retraction command is triggered. The retraction path is planned based on Bézier curves to ensure that the robotic arm retreats with a trajectory of minimum curvature.

[0053] Multi-machine collaborative collision avoidance strategies include: The central controller allocates the grasping area using a spatiotemporal constraint model to avoid overlapping workspaces for the robotic arm. A priority preemption mechanism dynamically calculates permission allocation weights based on task urgency and the robot's real-time load. When the paths of the two robots overlap, the system dynamically adjusts the trajectory using a repulsive force model: If the distance at the end of the robotic arm The system triggers collision avoidance trajectory replanning and dynamically adjusts the detour strategy based on global path weights.

[0054] In practice, multi-level fault detection and dynamic response include: the robot monitors operating status parameters in real time through built-in sensors, including remaining battery power. Motor temperature and communication delay When any parameter exceeds the safety threshold ( , or When the faulty robot is detected, the system triggers a three-level response mechanism: the primary response immediately freezes the robot's motion commands and broadcasts a fault code (including fault type, location, and load status); the intermediate response is based on the real-time status of neighboring robots (location). Remaining load and task priority Calculate the priority score for the support robot: To support the Euclidean distance between the robot and the fault point, This is the robot's maximum load capacity. This indicates the percentage of idle time. If the remaining load of a single robot is insufficient to take over the task, the system uses a Hungarian algorithm to break the task into multiple sub-tasks, ensuring that the global load deviation rate is below [a certain percentage]. The task breakdown formula is: in This represents the total weight of the faulty robot to be transferred. This represents the minimum number of splits.

[0055] Dynamic task migration and load balancing include: During the task transfer process, the system processes the queue of robots awaiting execution in descending order of urgency. Regarding weight For the goods, a parallel transfer strategy is adopted: the task is broken down into... There are 1 subtask, and the weight of each subtask is: The optimal allocation matrix is ​​generated by improving the Hungarian algorithm. The objective function is: in, As a task priority, To support the distance from the robot to the task point, the load verification phase calculates the actual load of each support robot: If overload occurs, the overloaded portion will be reallocated to ensure global load balancing. Global path coordination and conflict resolution include: After the task transfer is completed, the system updates the path weights in the spatiotemporal graph model. And adjust the support robot's motion trajectory based on a dynamic potential field model. When the number of robots in a certain channel At that time, its toll costs increased to The robot is guided to navigate around low-density areas; if the support robot's path overlaps with that of other robots, collision avoidance trajectory optimization is triggered, and the repulsion force calculation formula is: High-priority task robots are given priority access to fast passage paths, while low-priority robots enter a waiting queue to minimize the impact of path conflicts on overall efficiency.

[0056] Digital twin verification and self-learning optimization include: The system uses a digital twin engine to simulate the sorting process and pre-test the feasibility of task transfer schemes. If the simulated sorting efficiency... And the conflict rate If the condition is met, an execution command is issued; otherwise, the task splitting granularity and path weight coefficients are recalculated. Historical fault data is input into the reinforcement learning model to optimize the weight parameters of the scoring function (Equation 16). The objective function is: After 5000 iterations of training, the fault recovery time decreased from the initial... shortened to The success rate of task transfer has been increased to .

[0057] Digital twin and performance verification include: constructing a virtual-physical collaborative warehouse sorting simulation environment through the deep integration of digital twin technology and deep reinforcement learning, realizing full lifecycle verification and dynamic optimization of the sorting process. The following systematically elaborates on the practical application and implementation details of the technical solution from the dimensions of digital twin modeling and reinforcement learning training.

[0058] Digital twin modeling and virtual-real synchronization include: a digital twin engine building a high-precision virtual warehouse environment based on the Unity3D framework; and bidirectional real-time data synchronization between the physical and virtual layers through ROS middleware. This includes physical layer robot pose data. by Frequency updates are sent to the virtual environment, and cargo status (location, grabbing status) is triggered by incremental transmission via an event-driven model, with synchronization errors controlled within a specified range. Within this range, key parameters of the virtual environment (shelf layout, robot kinematics, and cargo weight distribution) are completely consistent with the physical layer, and it dynamically simulates the physical interaction behavior of the warehousing scenario. For example, during a surge in promotional orders, the system automatically generates peak order data streams (such as per hourly data). The system simulates the sorting process in a virtual environment to verify the robustness of the path planning algorithm under high load. Furthermore, the dynamic obstacle generation module can simulate unexpected scenarios such as shelf collapse and temporary stacking of goods to test the robot's dynamic obstacle avoidance and task redistribution capabilities. A virtual-real synchronization mechanism ensures that simulation results can be directly mapped to the physical layer, providing reliable pre-validation for actual deployment.

[0059] Reinforcement learning models and dynamic optimization include: deep reinforcement learning models that aim to maximize sorting efficiency and minimize conflict rate, and design state spaces. (Including robot position, battery level, load, and task queue length) and motion space (A = {allocation, preemption, waiting}). The reward function integrates task completion rate, conflict count, and energy consumption metrics. The first reward is for completing high-priority tasks on time (delay time). The exponential decay strengthens the time constraint, the second term suppresses path conflict frequency, and the last term penalizes energy consumption. The training process employs the Dual Deep Q-Network (DDQN) algorithm, where each term in the target network... Update once per step, the experience replay pool capacity is set to Exploration rate From initial value linear decay to Prioritized Experience Replay improves the learning efficiency of critical states. The sampling probability is calculated using the following formula: in For timing difference error .go through After several iterations of training, the task completion rate increased from the initial... Upgraded to The conflict rate has stabilized at The training results are mapped to the physical layer in real time through a digital twin environment, dynamically adjusting task allocation weights and path planning parameters. For example, when high-frequency path conflict areas are detected, the congestion penalty coefficient in the spatiotemporal graph model is automatically increased. This guides the robot to navigate around the low-density passageway.

[0060] This embodiment receives real-time order data, parses it to generate a task set containing the location of goods, target sorting area, and priority; through a distributed task allocation mechanism, it dynamically calculates the task matching degree by combining the state information of each robot, and uses a reinforcement learning model to optimize the global task allocation and generate a global task allocation result; based on the global allocation strategy, it models the warehouse environment as a spatiotemporal graph structure; it uses a pathfinding algorithm combined with the spatiotemporal graph structure to generate conflict-free paths and optimizes the trajectory by combining task time constraints; based on the target position in the conflict-free path, it identifies the goods through a vision system and uses a force-controlled robotic arm to complete the grasping and placement. This achieves the technical effect of improving sorting efficiency and adaptability, and reducing operation and maintenance costs.

[0061] Furthermore, embodiments of this application also propose a computer-readable storage medium storing a program for collaborative sorting of multiple Aelos robots in warehousing and logistics. When the program for collaborative sorting of multiple Aelos robots in warehousing and logistics is executed by a processor, it implements the steps of the method for collaborative sorting of multiple Aelos robots in warehousing and logistics as described above.

[0062] Reference Figure 5 , Figure 5 This is a structural block diagram of the first embodiment of the multi-Aelos robot collaborative sorting device for warehousing and logistics in this application.

[0063] like Figure 5 As shown in the embodiments of this application, the multi-Aelos robot collaborative sorting device for warehousing and logistics includes: Data acquisition module 10 is used to receive real-time order data and parse the real-time order data to generate a task set containing the location of goods, target sorting area and priority. The result allocation module 20 is used to dynamically calculate the task matching degree by combining the state information of each robot with the distributed task allocation mechanism, and to optimize the global task allocation by using a reinforcement learning model to generate the global task allocation result. The spatiotemporal graph module 30 is used to model the warehouse environment as a spatiotemporal graph structure based on a global allocation strategy; Optimization module 40 is used to generate a conflict-free path by combining the pathfinding algorithm with the spatiotemporal graph structure and optimize the trajectory by combining the task time limit constraints; The execution module 50 is used to identify the goods through a vision system based on the target position in the conflict-free path, and then combine the force-controlled robotic arm to complete the grasping and placement.

[0064] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solution of this application. In specific applications, those skilled in the art can make settings as needed, and this application does not impose any restrictions on this.

[0065] This embodiment receives real-time order data, parses it to generate a task set containing the location of goods, target sorting area, and priority; through a distributed task allocation mechanism, it dynamically calculates the task matching degree by combining the state information of each robot, and uses a reinforcement learning model to optimize the global task allocation and generate a global task allocation result; based on the global allocation strategy, it models the warehouse environment as a spatiotemporal graph structure; it uses a pathfinding algorithm combined with the spatiotemporal graph structure to generate conflict-free paths and optimizes the trajectory by combining task time constraints; based on the target position in the conflict-free path, it identifies the goods through a vision system and uses a force-controlled robotic arm to complete the grasping and placement. This achieves the technical effect of improving sorting efficiency and adaptability, and reducing operation and maintenance costs.

[0066] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this application. In practical applications, those skilled in the art can select some or all of it to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0067] In addition, for technical details not described in detail in this embodiment, please refer to the method for collaborative sorting of multiple Aelos robots for warehousing and logistics provided in any embodiment of this application, which will not be repeated here.

[0068] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0069] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0070] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application. The above are only preferred embodiments of this application and do not limit the patent scope of this application. All equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A multi-Aelos robot collaborative sorting method for warehousing and logistics, characterized in that, include: Receive real-time order data and parse the real-time order data to generate a task set that includes the location of goods, target sorting area and priority. Through a distributed task allocation mechanism, the task matching degree is dynamically calculated by combining the state information of each robot, and a reinforcement learning model is used to optimize the global task allocation and generate the global task allocation result. Based on the global allocation strategy, the warehouse environment is modeled as a spatiotemporal graph structure; The construction of the spatiotemporal graph structure includes: Add timestamps to key locations for dynamic node marking; Automatically increase passage costs based on real-time robot density in the passageway; Reserve dedicated passage periods for emergency tasks to achieve time window constraint integration; A pathfinding algorithm is used in conjunction with the spatiotemporal graph structure to generate a conflict-free path and introduce time window constraints. The trajectory is then optimized in conjunction with task time constraints. Based on the target location in the conflict-free path, the goods are identified by a vision system and then grasped and placed by a force-controlled robotic arm.

2. The method according to claim 1, characterized in that, The generated task set includes: Automatically identifies the weight, dimensions, and fragility characteristics of goods; Promotional orders are automatically marked as the highest priority and assigned to an emergency processing channel; The task queue is refreshed immediately when a new order arrives to ensure timeliness.

3. The method according to claim 1, characterized in that, Dynamically calculating task matching degree includes: A multi-dimensional status assessment is performed by integrating the robot's real-time location, remaining battery power, current load, and task urgency. Automatically adjust the weighting of the allocation strategy based on historical sorting data to achieve reinforcement learning optimization; Set up fast lanes for high-priority orders to skip the regular queuing process and enable competitive task allocation.

4. The method according to claim 1, characterized in that, Generating conflict-free paths includes: A hierarchical planning strategy is adopted, using spatiotemporal A at the global level. The algorithm generates the main path and avoids dynamic obstacles in real time at the local layer using a potential field model. Insert smooth curve segments into sharp turns to ensure stable robot steering and achieve smooth trajectory processing; Assign high-value order robots a direct fast lane to achieve priority passage.

5. The method according to claim 1, characterized in that, Grab and place operations include: Vision-force control dual closed loop: The vision system locates the coordinates of the goods at the millimeter level, and the force-controlled robotic arm automatically adjusts the gripping force according to the characteristics of the goods; Multi-robot collaborative collision avoidance: The central controller allocates staggered grasping time slots for adjacent robots; Special handling for fragile items: Automatically activates a buffer gripping mode for glass products, etc.

6. The method according to claim 1, characterized in that, The method also includes intelligent fault response: Multi-level early warning mechanism: A level three alarm is triggered when the battery level is below 20% or the temperature exceeds 70°C; Seamless task transfer: Automatically transfer tasks based on the location, idle status, and load capacity of nearby robots; Load balancing guarantee: Heavy tasks are intelligently split into multiple sub-tasks for distribution.

7. A multi-Aelos robot collaborative sorting device for warehousing and logistics, characterized in that, Performing the method as described in claim 1 includes: The data acquisition module is used to receive real-time order data and parse the real-time order data to generate a task set containing the location of goods, target sorting area and priority. The result allocation module is used to dynamically calculate the task matching degree by combining the state information of each robot with the distributed task allocation mechanism, and to optimize the global task allocation by using a reinforcement learning model to generate the global task allocation result. The spatiotemporal graph module is used to model the warehouse environment as a spatiotemporal graph structure based on a global allocation strategy; The optimization module is used to generate conflict-free paths by combining the pathfinding algorithm with the spatiotemporal graph structure and optimize the trajectory by combining the task time constraints. The execution module is used to identify the goods through a vision system based on the target position in the conflict-free path, and then combine it with a force-controlled robotic arm to complete the grasping and placement.

8. A computer device, characterized in that, The device includes a memory and a processor, wherein the processor, when executing computer instructions stored in the memory, performs the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 6.

Images