Robot task allocation method and device based on digital twinning

Abstract

The embodiment of the invention provides a robot task allocation method and device based on digital twinning, and the method comprises the steps: scanning a plurality of robots, and generating a virtual twinning model corresponding to the plurality of robots based on a scanning result; obtaining currently allocated tasks of the plurality of robots, and determining load evaluation information of the plurality of robots executing the currently allocated tasks; determining a target robot to which a task is to be allocated in the plurality of robots according to the load evaluation information of executing the currently allocated task by the plurality of robots; determining a collaborative operation scene, and performing task allocation simulation according to the virtual twin models corresponding to the plurality of robots and the collaborative operation scene to obtain a simulation allocation result; and performing task allocation on the target robot to be subjected to task allocation according to the simulation allocation result. Through efficient management of cooperation of multiple robots, the operation efficiency of a single robot and the overall efficiency of the system are improved.

Description

Technical Field

[0001] This invention relates to the field of robotics, and in particular to a robot task allocation method and apparatus based on digital twins. Background Technology

[0002] Robots are widely used in industrial settings and are continuously driving the intelligent upgrade of manufacturing, logistics, and healthcare. However, robots still have significant technical shortcomings in terms of operational efficiency. For example, in robot task allocation, fixed-division task allocation models are difficult to adapt to changes in load. In logistics palletizing scenarios, when assigning tasks to robots, some robots become overloaded due to a surge in orders, while others have low utilization rates. Due to the limitations of the system's task allocation, some robots have high loads and are prone to overloading, while others have low loads and low utilization rates, resulting in low overall operational efficiency for multiple robots and thus affecting the overall efficiency of the system. Summary of the Invention

[0003] To address the aforementioned problems, embodiments of the present invention disclose a robot task allocation method and apparatus based on digital twins.

[0004] In a first aspect, embodiments of the present invention provide a robot task allocation method based on digital twins, the method comprising:

[0005] Multiple robots are scanned, and virtual twin models corresponding to the multiple robots are generated based on the scan results;

[0006] Obtain the currently assigned tasks of the multiple robots, and determine the load assessment information of the multiple robots executing the currently assigned tasks;

[0007] Based on the load assessment information of the multiple robots performing the currently assigned tasks, determine the target robot among the multiple robots to be assigned tasks;

[0008] Determine the collaborative operation scenario, and based on the virtual twin models corresponding to the multiple robots, simulate task allocation according to the collaborative operation scenario to obtain the simulation allocation result;

[0009] Based on the simulation allocation results, the target robot to be assigned the task is assigned a task.

[0010] Optionally, the method further includes:

[0011] When an anomaly is detected among the plurality of robots, the abnormal robot among the plurality of robots is identified;

[0012] The operator performs virtual repair operations on the malfunctioning robot and evaluates these virtual repair operations.

[0013] Optionally, determining the load assessment information for the plurality of robots to perform the currently assigned task includes:

[0014] Determine the task saturation, motion complexity, and load change rate for each robot;

[0015] Based on the task saturation, motion complexity, and / or load change rate of each robot, determine the load assessment information for each robot to perform the currently assigned task.

[0016] Optionally, the collaborative operation scenario includes: a layered alternating scenario; the step of simulating task allocation according to the collaborative operation scenario based on the virtual twin models corresponding to the multiple robots, and obtaining the simulation allocation result, includes:

[0017] In the case where the collaborative operation scenario of the multiple robots is a layered and alternating scenario, the virtual position coordinates of each robot are determined, and the path movement simulation of the virtual twin model corresponding to the multiple robots is performed based on the virtual position coordinates of each robot to obtain the simulated operation area of ​​the target robot of the task to be assigned.

[0018] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0019] Tasks are assigned to the target robot based on the simulated working area of ​​the target robot to be assigned the task.

[0020] Optionally, the collaborative operation scenario includes: a priority rotation scenario; the step of simulating task allocation according to the collaborative operation scenario based on the virtual twin models corresponding to the multiple robots, and obtaining the simulation allocation result, includes:

[0021] In the case where the collaborative operation scenario of the multiple robots is a priority rotation scenario, the task level of the currently assigned tasks of each robot is obtained; based on the task level of the currently assigned tasks of each robot and the load assessment information of each robot, the task deviation between each robot is determined; based on the task deviation between each robot, the virtual twin model corresponding to each robot is used to simulate task rotation, and the simulation rotation result of the target robot of the task to be assigned is obtained.

[0022] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0023] Based on the simulation rotation results of the target robots for the tasks to be assigned, tasks are assigned to the target robots for the tasks to be assigned.

[0024] Optionally, the collaborative operation scenario includes: a capability complementarity scenario; the step of simulating task allocation according to the collaborative operation scenario based on the virtual twin models corresponding to the multiple robots, and obtaining the simulation allocation result, includes:

[0025] In the case where the collaborative operation scenario of the multiple robots is a scenario of complementary capabilities, the load capacity of each robot and the load parameters of the currently assigned tasks of each robot are obtained; based on the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, a task transition simulation is performed on the virtual twin model corresponding to the multiple robots to obtain the simulation transition result of the target robot of the task to be assigned.

[0026] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0027] The simulation transition results of the target robot to be assigned the task are used to assign the task to the target robot.

[0028] Optionally, determining the target robot to be assigned a task among the plurality of robots based on the load assessment information of the plurality of robots performing the currently assigned tasks includes:

[0029] Robots whose load assessment information is greater than or equal to a preset load threshold are selected as target robots for task assignment.

[0030] Optionally, the method further includes:

[0031] The load balance, total task duration, equipment wear and tear, and equipment energy consumption of the multiple robots are obtained; the load balance is determined based on the load standard deviation and average load of the multiple robots; the equipment wear and tear is determined based on the number of joint movements and load assessment information of the robots; and the equipment energy consumption is determined based on the motor power of the robots.

[0032] The tasks assigned to the multiple robots are optimized based on the load balancing, total task duration, equipment wear and tear, and equipment energy consumption.

[0033] Optionally, the step of scanning multiple robots and generating virtual twin models corresponding to the multiple robots based on the scanning results includes:

[0034] The structures of the multiple robots are subjected to three-dimensional scanning to obtain the three-dimensional data of the multiple robots respectively;

[0035] The 3D data of the multiple robots are preprocessed to obtain the initial virtual twin models corresponding to the multiple robots;

[0036] Obtain the state parameters of the multiple robots respectively;

[0037] When the state parameters of the multiple robots meet the preset conditions, the initial virtual twin models corresponding to the multiple robots are corrected to obtain the virtual twin models corresponding to the multiple robots respectively.

[0038] Optionally, the state parameters of the plurality of robots include the three-dimensional pose of the robot's end effector, the joint angles of the robot, the operating current of the robot, and the ambient temperature;

[0039] The preset conditions include: the robot's end-effector pose error is greater than a preset pose error; and / or, the robot's operating current exceeds a preset current range; and / or, the ambient temperature changes more than a preset temperature threshold within a preset time period.

[0040] The step of correcting the initial virtual twin models corresponding to the plurality of robots to obtain virtual twin models corresponding to the plurality of robots includes:

[0041] Obtain the current fluctuation influence coefficient of the multiple robots, and update the joint stiffness coefficient of the initial virtual twin model corresponding to the multiple robots based on the current fluctuation influence coefficient of the multiple robots.

[0042] The joint angular velocities of the multiple robots are obtained, and the joint angles of the initial virtual twin models corresponding to the multiple robots are updated based on the joint angular velocities of the multiple robots.

[0043] The change in ambient temperature is obtained, and the virtual twin model corresponding to the multiple robots is determined based on the end-effector pose error of the multiple robots, the joint stiffness coefficient after correction of the initial virtual twin model corresponding to the multiple robots, the joint angle after correction of the initial virtual twin model corresponding to the multiple robots, and the change in ambient temperature.

[0044] Optionally, the detection operator performs a virtual repair operation on the abnormal robot and evaluates the virtual repair operation, including:

[0045] The smoothness of the operation, the consistency of the operation steps, and the validity of the operation results of the virtual repair operation performed by the operator on the abnormal robot are determined; wherein, the smoothness of the operation is determined based on the operation time of the virtual repair operation and the preset time, the consistency of the operation steps is determined based on the consistency of the operation steps of the virtual repair operation with the preset process, and the validity of the operation results is determined based on the virtual damage ratio and the failure recovery time achievement rate of the virtual repair operation.

[0046] The operator's operation process is evaluated based on the smoothness of the virtual repair operation, the consistency of the operation steps, and the validity of the operation results.

[0047] Optionally, the method further includes:

[0048] The joint angular velocities of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot are obtained.

[0049] The virtual repair operation performed by the operator shall be stopped if the joint angular velocity of the abnormal robot is greater than or equal to a preset velocity threshold, and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold, and / or the end effector force of the abnormal robot is greater than or equal to a preset effector force threshold.

[0050] Secondly, embodiments of the present invention provide a robot task allocation device based on digital twins, the device comprising:

[0051] The virtual twin model generation module is used to scan multiple robots and generate virtual twin models corresponding to the multiple robots based on the scanning results.

[0052] The load assessment information determination module is used to obtain the currently assigned tasks of the multiple robots and determine the load assessment information of the multiple robots executing the currently assigned tasks.

[0053] The robot to be assigned module is used to determine the target robot among the multiple robots to be assigned the task based on the load assessment information of the multiple robots performing the currently assigned task;

[0054] The simulation allocation result determination module is used to determine the collaborative operation scenario, and to perform task allocation simulation according to the virtual twin models corresponding to the multiple robots and the collaborative operation scenario to obtain the simulation allocation result;

[0055] The simulated target task allocation module is used to allocate tasks to the target robot to be assigned tasks based on the simulated allocation results.

[0056] Optionally, the device further includes:

[0057] An abnormal robot detection module is used to identify the abnormal robot among the multiple robots when an abnormality is detected.

[0058] The virtual repair operation evaluation module is used to detect and evaluate the virtual repair operations performed by the operator on the abnormal robot.

[0059] Optionally, the load assessment information determination module includes:

[0060] The load assessment information acquisition submodule is used to determine the task saturation, motion complexity, and load change rate of each robot.

[0061] The load assessment information determination submodule is used to determine the load assessment information of each robot for performing the currently assigned task based on the task saturation, motion complexity and / or load change rate of each robot.

[0062] Optionally, the collaborative work scenario includes: a layered alternating scenario; the simulation allocation result determination module includes:

[0063] The simulated work area determination submodule is used to determine the virtual position coordinates of each robot when the collaborative work scenario of the multiple robots is a layered and alternating scenario, and to perform path movement simulation on the virtual twin model corresponding to the multiple robots based on the virtual position coordinates of each robot to obtain the simulated work area of ​​the target robot of the task to be assigned.

[0064] The simulated target task allocation module includes:

[0065] The first simulated target task allocation submodule is used to allocate tasks to the target robot based on the simulated working area of ​​the target robot to be assigned the task.

[0066] Optionally, the collaborative operation scenario includes: a priority rotation scenario; the simulated allocation result determination module includes:

[0067] The simulation rotation result determination submodule is used to obtain the task level of the currently assigned tasks of each robot when the collaborative operation scenario of the multiple robots is a priority rotation scenario; determine the task deviation between each robot based on the task level of the currently assigned tasks of each robot and the load assessment information of each robot; and perform task rotation simulation on the virtual twin model corresponding to each robot based on the task deviation between each robot to obtain the simulation rotation result of the target robot of the task to be assigned.

[0068] The simulated target task allocation module includes:

[0069] The second simulated target task allocation submodule is used to allocate tasks to the target robots of the tasks to be assigned based on the simulated rotation results of the target robots of the tasks to be assigned.

[0070] Optionally, the collaborative operation scenario includes: a capability complementarity scenario; the simulation allocation result determination module includes:

[0071] The simulation transition result determination submodule is used to obtain the load capacity of each robot and the load parameters of the currently assigned tasks of each robot when the collaborative operation scenario of the multiple robots is a complementary capability scenario; based on the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, the module performs task transition simulation on the virtual twin models corresponding to the multiple robots to obtain the simulation transition result of the target robot of the task to be assigned.

[0072] The simulated target task allocation module includes:

[0073] The third simulated target task allocation submodule is used to allocate tasks to the target robots based on the simulation transition results of the task to be assigned.

[0074] Optionally, the robot to be assigned determination module includes:

[0075] The robot to be assigned submodule is used to select robots whose load assessment information is greater than or equal to a preset load threshold from among the multiple robots as target robots for the task to be assigned.

[0076] Optionally, the device further includes:

[0077] The equipment information determination module is used to obtain the load balance, total task duration, equipment wear and tear, and equipment energy consumption of the multiple robots; the load balance is determined based on the load standard deviation and average load of the multiple robots; the equipment wear and tear is determined based on the number of joint movements and load assessment information of the robots; and the equipment energy consumption is determined based on the motor power of the robots.

[0078] The task allocation optimization module is used to optimize the tasks allocated to the multiple robots based on the load balance of the multiple robots, the total task duration, equipment wear and tear, and equipment energy consumption.

[0079] Optionally, the virtual twin model generation module includes:

[0080] A 3D data generation submodule is used to perform 3D scanning on the structure of the multiple robots to obtain 3D data of the multiple robots respectively;

[0081] The three-dimensional data preprocessing submodule is used to preprocess the three-dimensional data of the multiple robots to obtain the initial virtual twin models corresponding to the multiple robots respectively;

[0082] The state parameter acquisition submodule is used to acquire the state parameters of the multiple robots respectively;

[0083] The virtual twin model generation submodule is used to correct the initial virtual twin models corresponding to the multiple robots when the state parameters of the multiple robots meet the preset conditions, so as to obtain the virtual twin models corresponding to the multiple robots respectively.

[0084] Optionally, the state parameters of the plurality of robots include the three-dimensional pose of the robot's end effector, the joint angles of the robot, the operating current of the robot, and the ambient temperature;

[0085] The preset conditions include: the robot's end-effector pose error is greater than a preset pose error; and / or, the robot's operating current exceeds a preset current range; and / or, the ambient temperature changes more than a preset temperature threshold within a preset time period.

[0086] The virtual twin model generation submodule includes:

[0087] The joint stiffness coefficient update unit is used to obtain the current fluctuation influence coefficient of the multiple robots, and update the joint stiffness coefficient of the initial virtual twin model corresponding to the multiple robots according to the current fluctuation influence coefficient of the multiple robots.

[0088] The joint angle update unit is used to obtain the joint angular velocities of the multiple robots and update the joint angles of the initial virtual twin models corresponding to the multiple robots based on the joint angular velocities of the multiple robots.

[0089] The virtual twin model generation unit is used to acquire the change in ambient temperature and determine the virtual twin model corresponding to the multiple robots based on the end-effector pose error of the multiple robots, the joint stiffness coefficient after correction of the initial virtual twin model corresponding to the multiple robots, the joint angle after correction of the initial virtual twin model corresponding to the multiple robots, and the change in ambient temperature.

[0090] Optionally, the virtual repair operation evaluation module includes:

[0091] The virtual repair operation determination submodule is used to determine the smoothness of the operation, the consistency of the operation steps, and the validity of the operation results of the virtual repair operation performed by the operator on the abnormal robot; wherein, the smoothness of the operation is determined based on the operation duration of the virtual repair operation and the preset duration, the consistency of the operation steps is determined based on the consistency of the operation steps of the virtual repair operation with the preset process, and the validity of the operation results is determined based on the virtual damage ratio and the failure recovery time achievement rate of the virtual repair operation;

[0092] The virtual repair operation evaluation submodule is used to evaluate the operator's operation process based on the smoothness of the operation, the consistency of the operation steps, and the validity of the operation results.

[0093] Optionally, the device further includes:

[0094] An abnormal robot information acquisition module is used to acquire the joint angular velocity of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot.

[0095] The virtual repair operation stop module is used to stop the operator's virtual repair operation when the joint angular velocity of the abnormal robot is greater than or equal to a preset speed threshold, and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold, and / or the end effector force of the abnormal robot is greater than or equal to a preset effector force threshold.

[0096] Thirdly, the present invention discloses an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described robot task allocation method based on digital twin.

[0097] Fourthly, the present invention discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described robot task allocation method based on digital twins.

[0098] The embodiments of the present invention have the following advantages:

[0099] This invention can scan multiple robots, generate virtual twin models corresponding to the robots based on the scanning results, and ensure accurate mapping of robot models by using digital twin technology to generate virtual models corresponding to robots. It acquires the currently assigned tasks of the multiple robots and determines the load assessment information of the multiple robots executing the currently assigned tasks. Based on the load assessment information of the multiple robots executing the currently assigned tasks, it determines the target robot among the multiple robots to be assigned tasks. By assessing the task load of each robot, it can prioritize selecting a lightly loaded robot to undertake tasks, balancing the load and extending equipment life. It determines collaborative operation scenarios, and performs task allocation simulation according to the virtual twin models corresponding to the multiple robots, obtaining simulated allocation results. It can perform corresponding task allocation simulations for different collaborative operation scenarios, improving system flexibility. Moreover, simulating robot operation through virtual twin models can avoid actual conflicts and improve system stability. Based on the simulated allocation results, it assigns tasks to the target robots to be assigned tasks. By dynamically adjusting the task allocation, it can reduce the operation time of a single robot, improve load balancing, and increase robot operation efficiency. The embodiments of the present invention can efficiently manage the collaboration of multiple robots through real-time data acquisition, load balancing, virtual simulation and dynamic task scheduling, thereby improving the working efficiency of individual robots and the overall efficiency of the system. Attached Figure Description

[0100] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0101] Figure 1 This is a flowchart illustrating the steps of a robot task allocation method based on digital twins according to an embodiment of the present invention.

[0102] Figure 2 This is a flowchart illustrating the steps of another robot task allocation method based on digital twins according to an embodiment of the present invention.

[0103] Figure 3 This is a logic diagram of the exclusive twin initialization and dynamic calibration in an embodiment of the present invention;

[0104] Figure 4 This is a logic diagram of dynamic task allocation and optimization in an embodiment of the present invention;

[0105] Figure 5 This is a logic diagram of virtual-real linkage fault handling and teaching evaluation according to an embodiment of the present invention;

[0106] Figure 6 This is a logical diagram of the multi-technology fusion system architecture according to an embodiment of the present invention;

[0107] Figure 7 This is a structural block diagram of a robot task allocation device based on digital twin according to an embodiment of the present invention. Detailed Implementation

[0108] This invention proposes a robot task allocation method based on digital twins, aiming to reduce the operation time of a single robot, improve load balancing, and enhance system performance, security, and economy. To achieve this goal, embodiments of this invention use digital twin technology to map the physical entities of multiple robots into a virtual model, dynamically allocating tasks according to the collaborative operation scenario. This allows for efficient management of multi-robot collaboration, improving the operational efficiency of individual robots and the overall efficiency of the system.

[0109] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0110] Reference Figure 1 The diagram illustrates a flowchart of a robot task allocation method based on digital twins according to an embodiment of the present invention. The method may specifically include the following steps:

[0111] Step 101: Scan multiple robots and generate virtual twin models corresponding to the multiple robots based on the scanning results;

[0112] In this embodiment of the invention, multiple robots can be scanned, and virtual twin models corresponding to the multiple robots can be generated based on the scan results. Through data acquisition in physical space, a high-fidelity, real-time synchronized virtual twin model with the real robot is constructed. Specifically, high-density scanning (point cloud spacing ≤ 0.1 mm) can be used for key structures of the robot body (such as joint connectors and end effectors), while conventional density scanning (point cloud spacing ≤ 0.5 mm) is used for non-critical areas, ensuring the integrity of feature details while avoiding data redundancy. After scanning, a three-dimensional model is generated through point cloud denoising, stitching, and surface reconstruction.

[0113] Step 102: Obtain the currently assigned tasks of the multiple robots and determine the load assessment information of the multiple robots executing the currently assigned tasks;

[0114] In this embodiment of the invention, load assessment information for multiple robots performing currently assigned tasks can be determined. The load assessment information L = w1 × workload percentage + w2 × motion complexity + w3 × load change rate. Workload percentage = current executed task volume / rated daily task volume (reflecting task saturation); motion complexity = number of trajectory inflection points × average curvature (quantifying motion difficulty); load change rate = load fluctuation amplitude per unit time (reflecting load stability); weights w1, w2, and w3 are dynamically adjusted according to the real-time scenario (e.g., w2 increases to 0.4 for precision operations, and w1 increases to 0.5 for batch handling), and satisfy w1 + w2 + w3 = 1.

[0115] Step 103: Based on the load assessment information of the multiple robots executing the currently assigned tasks, determine the target robot among the multiple robots to be assigned tasks;

[0116] In this embodiment of the invention, the target robot for the assigned task can be determined from among multiple robots based on the load assessment information of multiple robots performing currently assigned tasks. For example, a robot with load assessment information L ≥ an adjustment threshold (e.g., 0.85) can be selected as the target robot for the assigned task.

[0117] Step 104: Determine the collaborative operation scenario; based on the virtual twin models corresponding to the multiple robots, perform task allocation simulation according to the collaborative operation scenario to obtain the simulation allocation result.

[0118] In this embodiment of the invention, a collaborative work scenario can be the collaborative logic and behavioral pattern followed by multiple robots when jointly completing a certain type of task or goal. It defines the robot's role, interaction method, spatiotemporal constraints, and collaboration rules. The collaborative work scenario in this embodiment includes three dynamically switchable work modes: a hierarchical alternation mode, a priority rotation mode, and a capability complementarity mode, ensuring load balancing and maximizing efficiency.

[0119] Specifically, in a scenario where multiple robots are working collaboratively in a layered and alternating manner, the virtual twin models of multiple robots can be used to simulate path movement based on the virtual position coordinates of each robot, thereby obtaining the simulated working area of ​​the target robot to be assigned the task.

[0120] In a priority-based rotation scenario where multiple robots are working collaboratively, the task deviation between robots can be determined based on the task level of each robot's currently assigned task and the load assessment information of each robot. Based on the task deviation between robots, task rotation simulation is performed on the virtual twin model corresponding to each robot to obtain the simulated rotation result of the target robot to be assigned the task.

[0121] In collaborative operation scenarios where multiple robots have complementary capabilities, task transition simulations can be performed on virtual twin models corresponding to multiple robots based on the load capacity of each robot and the load parameters of the currently assigned tasks, thereby obtaining the simulated transition results of the target robot for the task to be assigned.

[0122] Step 105: Based on the simulation allocation results, assign tasks to the target robot to be assigned tasks.

[0123] This invention can scan multiple robots, generate virtual twin models corresponding to the robots based on the scanning results, and ensure accurate mapping of robot models by using digital twin technology to generate virtual models corresponding to robots. It acquires the currently assigned tasks of the multiple robots and determines the load assessment information of the multiple robots executing the currently assigned tasks. Based on the load assessment information of the multiple robots executing the currently assigned tasks, it determines the target robot among the multiple robots to be assigned tasks. By assessing the task load of each robot, it can prioritize selecting a lightly loaded robot to undertake tasks, balancing the load and extending equipment life. It determines collaborative operation scenarios, and performs task allocation simulation according to the virtual twin models corresponding to the multiple robots, obtaining simulated allocation results. It can perform corresponding task allocation simulations for different collaborative operation scenarios, improving system flexibility. Moreover, simulating robot operation through virtual twin models can avoid actual conflicts and improve system stability. Based on the simulated allocation results, it assigns tasks to the target robots to be assigned tasks. By dynamically adjusting the task allocation, it can reduce the operation time of a single robot, improve load balancing, and increase robot operation efficiency. The embodiments of the present invention can efficiently manage the collaboration of multiple robots through real-time data acquisition, load balancing, virtual simulation and dynamic task scheduling, thereby improving the working efficiency of individual robots and the overall efficiency of the system.

[0124] Reference Figure 2 The diagram illustrates a flowchart of another robot task allocation method based on digital twins according to an embodiment of the present invention. The method may specifically include the following steps:

[0125] Step 201: Scan multiple robots and generate virtual twin models corresponding to the multiple robots based on the scanning results;

[0126] In one embodiment, the step of scanning multiple robots and generating virtual twin models corresponding to the multiple robots based on the scanning results may further include the following sub-steps:

[0127] Sub-step S11: Perform a three-dimensional scan on the structure of the multiple robots to obtain the three-dimensional data of the multiple robots respectively;

[0128] Sub-step S12: Preprocess the 3D data of the multiple robots to obtain the initial virtual twin models corresponding to the multiple robots respectively;

[0129] Sub-step S13: Obtain the state parameters of the plurality of robots respectively;

[0130] Sub-step S14: When the state parameters of the multiple robots meet the preset conditions, the initial virtual twin models corresponding to the multiple robots are corrected to obtain the virtual twin models corresponding to the multiple robots respectively.

[0131] This invention, by correcting the initial virtual twin models corresponding to multiple robots when the state parameters of multiple robots meet preset conditions, can make the virtual twin models more closely resemble the behavior of real robots. This allows for task allocation to multiple robots based on accurate virtual twin models, thereby improving the reliability of the system.

[0132] In one embodiment, the state parameters of the plurality of robots include the three-dimensional pose of the robot's end effector, the joint angles of the robot, the operating current of the robot, and the ambient temperature.

[0133] The preset conditions include: the robot's end-effector pose error is greater than a preset pose error; and / or, the robot's operating current exceeds a preset current range; and / or, the ambient temperature changes more than a preset temperature threshold within a preset time period.

[0134] The step of correcting the initial virtual twin models corresponding to the plurality of robots to obtain virtual twin models corresponding to the plurality of robots includes: obtaining the current fluctuation influence coefficient of the plurality of robots, and updating the joint stiffness coefficient of the initial virtual twin model corresponding to the plurality of robots according to the current fluctuation influence coefficient of the plurality of robots; obtaining the joint angular velocity of the plurality of robots, and updating the joint angle of the initial virtual twin model corresponding to the plurality of robots according to the joint angular velocity of the plurality of robots; obtaining the change in ambient temperature, and determining the virtual twin model corresponding to the plurality of robots according to the end-effector pose error of the plurality of robots, the corrected joint stiffness coefficient of the initial virtual twin model corresponding to the plurality of robots, the corrected joint angle of the initial virtual twin model corresponding to the plurality of robots, and the change in ambient temperature.

[0135] In this embodiment of the invention, high-density scanning can be used for key structures of the robot body (such as joint connectors and end effectors), while conventional density scanning is used for non-critical areas. This ensures the integrity of feature details while avoiding data redundancy. Then, the 3D data of multiple robots are preprocessed (point cloud denoising, stitching, and surface reconstruction) to obtain initial virtual twin models corresponding to multiple robots. The surface geometric error is controlled by ΔG = the spatial distance between the corresponding feature points of the physical entity and the virtual model, requiring ΔG ≤ G0 (G0 is set according to the industry scenario, for example, G0 = 0.2mm for precision assembly scenarios and G0 = 0.5mm for handling scenarios).

[0136] After importing into the simulation software, the robot's physical parameters can be matched using the formula "material density = base density × structural correction factor". The structural correction factor can be dynamically adjusted based on the manufacturing process of the physical components (such as casting or welding) to ensure that the mass, stiffness, and other characteristics of the virtual twin model are consistent with the robot entity, providing an accurate foundation for subsequent dynamic simulations.

[0137] To address the parameter drift issue during long-term robot operation, this invention constructs a dynamic calibration system encompassing the entire process of "perception-correction-verification." Specifically, multiple robot state parameters are acquired, including the robot's end-effector 3D pose, joint angles, operating current, and ambient temperature. When the robot's end-effector pose error exceeds a preset pose error; and / or the robot's operating current exceeds a preset current range; and / or the ambient temperature change within a preset time exceeds a preset temperature threshold, the initial virtual twin models corresponding to the multiple robots are corrected.

[0138] This invention can acquire the three-dimensional pose of the end effector in real time using a laser tracker, and simultaneously obtain the angle values ​​of the joint encoder, the motor operating current, and the ambient temperature (accuracy ±0.5℃) to form a multi-dimensional state dataset. The calibration process is automatically initiated when any of the following conditions are met to ensure that the error is controllable: end effector pose error ΔP = spatial distance between predicted pose and actual pose > P0 (P0 is a sub-millimeter threshold); joint current fluctuation ΔI (standard deviation of multiple consecutive sampling periods) > set threshold (e.g., 5% of rated current); ambient temperature change ΔT (temperature difference within 10 minutes) > sensitive threshold (e.g., 5℃).

[0139] The parameter chain collaborative correction logic is as follows: Based on the mapping relationship between current fluctuations and load, the joint stiffness coefficient K' = K0 × (1 + current fluctuation influence coefficient) is dynamically updated. For example, when the current fluctuation increases by 10%, the stiffness coefficient is reduced by 3% to compensate for elastic deformation. For scenarios where the robot's end effector linear velocity is >1m / s, centrifugal force correction is introduced. The angle correction amount is positively correlated with the square of the angular velocity (e.g., if the angular velocity increases by 20%, the correction amount increases by 4%). Combining the end effector pose error ΔP, the corrected stiffness coefficient K', and the temperature change ΔT, the initial virtual twin models corresponding to multiple robots are corrected to ensure that the final pseudo-twin model adapts to the robot's physical state in real time.

[0140] Step 202: Obtain the currently assigned tasks of the multiple robots and determine the load assessment information of the multiple robots executing the currently assigned tasks;

[0141] In one embodiment, the step of determining the load assessment information for the plurality of robots to perform the currently assigned task may further include the following sub-steps:

[0142] Sub-step S21: Determine the task saturation, motion complexity, and load change rate of each robot;

[0143] Sub-step S22: Based on the task saturation, motion complexity and / or load change rate of each robot, determine the load assessment information for each robot to perform the currently assigned task.

[0144] In this embodiment of the invention, the load assessment information for each robot performing its currently assigned tasks can be determined based on the task saturation, motion complexity, and / or load change rate of each robot. Specifically, the load assessment information L = w1 × workload percentage + w2 × motion complexity + w3 × load change rate. Wherein, workload percentage = current executed task volume / rated daily task volume (reflecting task saturation); motion complexity = number of trajectory inflection points × average curvature (quantifying motion difficulty); load change rate = load fluctuation amplitude per unit time (reflecting load stability); the weights w1, w2, and w3 are dynamically adjusted according to the real-time scenario (e.g., w2 increases to 0.4 for precision operations, and w1 increases to 0.5 for batch handling), and satisfy w1 + w2 + w3 = 1.

[0145] This invention provides a multi-dimensional robot load assessment method based on the task saturation, motion complexity, and / or load change rate of each robot. This method can identify robots that are about to be overloaded in advance, proactively divert tasks, and provide a high-quality decision-making basis for task allocation and load balancing.

[0146] Step 203: Based on the load assessment information of the multiple robots executing the currently assigned tasks, determine the target robot among the multiple robots to be assigned tasks;

[0147] In one embodiment, the step of determining the target robot among the plurality of robots to be assigned a task based on the load assessment information of the plurality of robots performing the currently assigned task may further include the following sub-steps:

[0148] Sub-step S31: Select the robot whose load assessment information is greater than or equal to a preset load threshold from among the multiple robots as the target robot to be assigned a task.

[0149] In this embodiment of the invention, robots whose load assessment information is greater than or equal to a preset load threshold can be selected as target robots for task assignment. For example, the preset load threshold can be 0.85. Those skilled in the art can set the preset load threshold to other appropriate values ​​according to the concept of this invention, and this invention does not limit this.

[0150] If the robot's load assessment information is greater than or equal to the preset load threshold, it indicates that the robot is in a high load, overloaded, or about to be overloaded state. By comparing the load assessment information with the preset load threshold, new tasks can be avoided from being assigned to overloaded robots, thus achieving intelligent scheduling that "prevents problems before they arise".

[0151] Step 204: Determine the collaborative operation scenario; based on the virtual twin models corresponding to the multiple robots, perform task allocation simulation according to the collaborative operation scenario to obtain the simulation allocation result.

[0152] In one embodiment, the collaborative operation scenario includes: a hierarchical alternation scenario, a priority rotation scenario, and a capability complementarity scenario; the step of simulating task allocation according to the virtual twin models corresponding to the multiple robots and obtaining the simulation allocation result may further include the following sub-steps:

[0153] Sub-step S41: In the case that the collaborative operation scenario of the multiple robots is a layered alternating scenario, determine the virtual position coordinates of each robot, and perform path movement simulation on the virtual twin model corresponding to the multiple robots based on the virtual position coordinates of each robot to obtain the simulated operation area of ​​the target robot of the task to be assigned.

[0154] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0155] Tasks are assigned to the target robot based on the simulated working area of ​​the target robot to be assigned the task.

[0156] In this embodiment of the invention, when the collaborative operation scenario of multiple robots is a layered and alternating scenario, the virtual twin models corresponding to multiple robots can be used to simulate path movement based on the virtual position coordinates of each robot to obtain the simulated operation area of ​​the target robot to be assigned the task.

[0157] For example, the workspace can be divided into a three-dimensional grid (e.g., 1m × 1m × 0.5m). Robot A is responsible for the work in the odd-numbered grid areas, and Robot B is responsible for the even-numbered grid areas. The switching interval is achieved through real-time coordinate comparison to avoid path conflicts. Layered alternation scenarios can be applied to large-area flat operations (such as warehouse sorting), thereby reducing the cross-movement distance of the robots.

[0158] Sub-step S42: In the case that the collaborative operation scenario of the multiple robots is a priority rotation scenario, obtain the task level of the currently assigned task of each robot; determine the task deviation between each robot based on the task level of the currently assigned task of each robot and the load assessment information of each robot; and perform task rotation simulation on the virtual twin model corresponding to each robot based on the task deviation between each robot to obtain the simulation rotation result of the target robot of the task to be assigned.

[0159] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0160] Based on the simulation rotation results of the target robots for the tasks to be assigned, tasks are assigned to the target robots for the tasks to be assigned.

[0161] In an embodiment of the present invention,

[0162] In a priority-based rotation scenario where multiple robots are working collaboratively, the task deviation between robots can be determined based on the task level of each robot's currently assigned task and the load assessment information of each robot. Based on the task deviation between robots, task rotation simulation is performed on the virtual twin model corresponding to each robot to obtain the simulated rotation result of the target robot to be assigned the task.

[0163] For example, tasks can be categorized by "urgency (E) - complexity (C)" (E and C are both levels 1-5), and executed in rotation among robots according to the rule "E1C1→E2C2→..." when under high load. For instance, when robot A's load assessment information L=0.9, tasks of robot A with E=3 and C=4 can be rotated to robot B with load assessment information L=0.6, ensuring that the ∑(E×C) task deviation between the two robots is ≤10%, thus avoiding overload of a single device.

[0164] Sub-step S43: In the case that the collaborative operation scenario of the multiple robots is a capability complementarity scenario, obtain the load capacity of each robot and the load parameters of the currently assigned tasks of each robot; according to the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, perform task transition simulation on the virtual twin model corresponding to the multiple robots to obtain the simulation transition result of the target robot of the task to be assigned.

[0165] The step of assigning tasks to the target robot based on the simulation assignment results includes:

[0166] The simulation transition results of the target robot to be assigned the task are used to assign the task to the target robot.

[0167] In this embodiment of the invention, when the collaborative operation scenario of multiple robots is a scenario of complementary capabilities, task transition simulation can be performed on the virtual twin models corresponding to multiple robots based on the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, so as to obtain the simulation transition result of the target robot to be assigned tasks.

[0168] For example, task transition can be performed based on the robot's load capacity (maximum load, maximum speed, positioning accuracy, etc.) and the load parameters of the currently assigned task. When the load parameters (e.g., the required load of 20kg) exceed the current robot's load capacity (e.g., the maximum load of 15kg), the task can be assigned to a suitable robot (e.g., robot C with a maximum load of 30kg). The task transition path can be planned according to "starting position + adaptive speed integral" to ensure path smoothness, thereby breaking through the limitations of fixed robot division of labor, shortening the continuous operation time of a single robot, and improving the collaborative efficiency of multiple robots.

[0169] Step 205: Based on the simulation allocation results, assign tasks to the target robots for the tasks to be assigned;

[0170] In this embodiment of the invention, corresponding task allocation simulations can be performed for different collaborative operation scenarios. Based on the simulation allocation results, the target robot to which the task is to be assigned is assigned a task. By dynamically adjusting the task allocation, the operation time of a single robot can be reduced, the load balance can be improved, and the operation efficiency of the robot can be increased.

[0171] Step 206: Obtain the load balance, total task duration, equipment wear and tear, and equipment energy consumption of the multiple robots; the load balance is determined based on the load standard deviation and average load of the multiple robots; the equipment wear is determined based on the number of joint movements and load assessment information of the robots; the equipment energy consumption is determined based on the motor power of the robots.

[0172] Step 207: Optimize the tasks assigned to the multiple robots based on their load balancing, total task duration, equipment wear and tear, and equipment energy consumption.

[0173] In this embodiment of the invention, to avoid efficiency degradation caused by sudden load increases, a proactive control strategy of "prediction-pre-allocation-optimization" can be implemented. The current load curve is fitted using a sliding window algorithm, and the load curve L(t) for the next 5 minutes is predicted by combining historical data from the same period. When L(t) is predicted to be ≥ a preset load threshold, 20%–40% of the non-emergency tasks of the current robot can be pre-allocated to idle robots, with the spatial overlap rate of the pre-allocated trajectories ≤ 5% (verified by a collision detection algorithm). The core of the optimization is a comprehensive objective function F = 0.4 × load balance + 0.3 × total task duration + 0.2 × equipment loss + 0.1 × equipment energy consumption. Wherein, load balance = 1 - robot load standard deviation / average load; equipment loss is positively correlated with the number of joint movements and load duration; equipment energy consumption is calculated as the integral of motor power. In emergency scenarios (such as fault repair), the weight of the total task duration is increased to 0.5, while in normal scenarios, the weight of the load balance is kept at 0.4. The optimal allocation scheme is solved using a particle swarm optimization algorithm.

[0174] This invention improves overall system stability by using load prediction, dynamic task pre-allocation, multi-objective optimization, and intelligent algorithm solutions to anticipate load trends, proactively pre-allocate tasks, avoid overload, and prevent overload.

[0175] Step 208: When an anomaly is detected in the plurality of robots, the abnormal robot among the plurality of robots is identified;

[0176] Step 209: Detect the virtual repair operation performed by the operator on the abnormal robot, and evaluate the virtual repair operation;

[0177] In one embodiment, the step of detecting the virtual repair operation performed by the operator on the abnormal robot and evaluating the virtual repair operation may further include the following sub-steps:

[0178] Sub-step S51: Determine the smoothness of the operation, the consistency of the operation steps, and the validity of the operation result of the virtual repair operation performed by the operator on the abnormal robot; wherein, the smoothness of the operation is determined based on the operation time of the virtual repair operation and the preset time, the consistency of the operation steps is determined based on the consistency of the operation steps of the virtual repair operation with the preset process, and the validity of the operation result is determined based on the virtual damage ratio and the failure recovery time achievement rate of the virtual repair operation;

[0179] Sub-step S52: Evaluate the operator's operation process based on the smoothness of the virtual repair operation, the consistency of the operation steps, and the validity of the operation results.

[0180] In this embodiment of the invention, to achieve accurate reproduction and early warning of faults, operators can be instructed to perform virtual repair operations on the abnormal robot. Specifically, when an encoder signal jump is detected, the virtual twin model can synchronously simulate changes in multiple physical fields: temperature field, T(t) = T0 + temperature rise coefficient × fault duration (the temperature rise coefficient is set according to the degree of current overload, such as 0.5℃ / s when the current exceeds the rated value by 20%); vibration characteristics, the vibration frequency increases exponentially from the normal 10Hz as the fault intensifies (such as reaching 50Hz 5s after the fault); current waveform, superimposed with random interference signals, the interference intensity increases linearly with the fault duration (such as the interference amplitude reaching 15% of the rated current after 10s).

[0181] Additionally, a diffusion threshold can be set (e.g., a temperature of 60°C or a vibration acceleration of 5 m / s²). 2 Once the threshold is exceeded, the probability of fault propagation is positively correlated with the current load assessment information (L value) and the duration. The higher the load and the longer the duration, the higher the probability of propagation is (10%) every 10 seconds. The amount of physical damage (D) can be quantified by the time integral of temperature and vibration, providing a basis for assessing the consequences of the fault.

[0182] This invention can construct a closed-loop interactive system of "physical operation - virtual feedback - physical adjustment," supporting safe teaching and remote operation. Operators can wear motion capture devices (such as inertial sensors), and their hand postures (angles, displacements) are wirelessly transmitted and mapped to the virtual robot, ensuring consistency between virtual repair operations and actual operations. When the virtual robot, or virtual twin model, enters a dangerous critical zone (e.g., ≤10cm from an obstacle), the system can generate a warning trajectory and trigger multiple layers on the physical end, such as increasing the buzzer frequency as the distance to the danger zone decreases (e.g., reaching 2kHz at 5cm); providing voice prompts for specific angle adjustments; and displaying simulated collision consequences (e.g., component deformation animations) in the virtual scene, enhancing the operator's risk awareness.

[0183] After the repair operation is completed, the virtual repair operation performed by the operator on the malfunctioning robot can be quantitatively scored. For example, the operation fluency S1 is scored based on the percentage of time-out operations (exceeding the standard time) (100 points for no timeout, 80 points for a 20% timeout percentage); the operation step conformity S2 is the comparison of the operation steps with the standard process (100 points for complete conformity, 20 points deducted per key step deviation); the operation result effectiveness S3 is scored by combining the virtual damage percentage (Dactual / Dmaximum) and the fault recovery time compliance rate (100 points for damage ≤10% and time compliance). The comprehensive score S = S1 × 0.3 + S2 × 0.4 + S3 × 0.3, and personalized improvement plans are output based on the score (e.g., strengthening standard process teaching for those with low S2 scores).

[0184] This invention, through quantitative scoring of virtual repair operations performed by operators on malfunctioning robots in a virtual environment, not only improves the quality of maintenance training and fault response, but also provides measurable and optimizable data support for operation and maintenance management.

[0185] Step 210: Obtain the joint angular velocity of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot;

[0186] Step 211: If the joint angular velocity of the abnormal robot is greater than or equal to a preset velocity threshold, and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold, and / or the end effector force of the abnormal robot is greater than or equal to a preset effector force threshold, the virtual repair operation of the operator shall be stopped.

[0187] In this embodiment of the invention, to ensure operational safety and teaching effectiveness, a safety assessment can be performed based on the joint angular velocity of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot. For example, when the joint angular velocity of the abnormal robot is greater than or equal to a preset speed threshold (e.g., 180° / s), and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold (e.g., 5 cm), and / or the end effector force of the abnormal robot is greater than or equal to a preset operating force threshold (e.g., 50 N), the virtual repair operation can be frozen, a correction scheme (e.g., deceleration command, path offset suggestion) can be generated, and the scheme can be unlocked after the adjustment meets the requirements.

[0188] This invention provides real-time safety assessments of key parameters of abnormal robots during virtual operations, such as joint angular velocity, distance from obstacles, and end effector force. This allows for early identification of potential dangers and proactive intervention, thereby enhancing the system's safety boundaries and providing scientific feedback for teaching and training.

[0189] To enable those skilled in the art to better understand the embodiments of the present invention, the following describes... Figure 3-6 The embodiments of the present invention will be described below. (Refer to...) Figure 3 The diagram illustrates the logic of dedicated twin initialization and dynamic calibration provided in an embodiment of the present invention. The method for dedicated twin initialization and dynamic calibration may specifically include the following steps:

[0190] Step 301: Perform a 3D scan on the physical robot to generate an initial 3D model and match material parameters;

[0191] Step 302: Calculate the distance error between the physical and virtual model feature points. If the error is less than or equal to a set threshold, proceed to the next step; otherwise, rescan and remodel.

[0192] Step 303: Real-time acquisition of multi-source data, including end-effector pose, joint current, ambient temperature, etc.

[0193] Step 304: When the end-effector pose error exceeds the sub-millimeter threshold, the current fluctuation exceeds 5%, or the temperature change exceeds 5°C, parameter correction is initiated.

[0194] Step 305: Dynamically adjust joint stiffness and kinematic parameters, and combine with LSTM model to predict error trends;

[0195] Step 306: After correction, verify the accuracy through 3 sets of trajectories. If the accuracy is met, the initialization is completed; otherwise, repeat the correction.

[0196] In this embodiment of the invention, to achieve high-precision mapping between the virtual model and the physical entity, a modeling strategy of "dynamic scanning + parameter adaptation" is adopted. Specifically, high-density scanning (point cloud spacing ≤ 0.1 mm) is used for key structures of the robot body (such as joint connectors and end effectors), while conventional density scanning (point cloud spacing ≤ 0.5 mm) is used for non-critical areas, ensuring the integrity of feature details while avoiding data redundancy. After scanning, a 3D model is generated through point cloud denoising, stitching, and surface reconstruction. Its surface geometric error is quantified and controlled by ΔG = the spatial distance between the corresponding feature points of the physical entity and the virtual model, requiring ΔG ≤ G0 (G0 is set according to the industry scenario, for example, G0 = 0.2 mm for precision assembly scenarios and G0 = 0.5 mm for handling scenarios).

[0197] After importing into the simulation software, the physical parameters are matched according to the formula "Material density = Base density × Structural correction factor". The structural correction factor is dynamically adjusted according to the processing technology of the physical component (such as casting or welding) (for example, the correction factor is 1.05 for welded components and 1.02 for cast components) to ensure that the mass, stiffness and other characteristics of the virtual model are consistent with the physical component, providing an accurate basis for subsequent dynamic simulation.

[0198] To address the parameter drift issue during long-term robot operation, a dynamic calibration system encompassing the entire process of "perception-correction-verification" is constructed. The specific implementation steps are as follows: The three-dimensional pose of the end effector is acquired in real time using a laser tracker (sampling frequency 1kHz), and the angle values ​​of the joint encoder (resolution 0.001°), motor operating current (sampling frequency 500Hz), and ambient temperature (accuracy ±0.5℃) are simultaneously obtained to form a multi-dimensional state dataset.

[0199] The calibration process will be automatically initiated when any of the following conditions are met to ensure that the error is controllable: End-effector pose error ΔP = spatial distance between predicted pose and actual pose > P0 (P0 is a sub-millimeter threshold, such as 0.1mm); Joint current fluctuation ΔI (standard deviation of 10 consecutive sampling periods) > set threshold (such as 5% of rated current); Ambient temperature change ΔT (temperature difference within 10 minutes) > sensitivity threshold (such as 5℃).

[0200] The parameter chain collaborative correction logic is as follows: Based on the mapping relationship between current fluctuation and load, the joint stiffness coefficient K' = K0 × (1 + current fluctuation influence coefficient) is dynamically updated. For example, when the current fluctuation increases by 10%, the stiffness coefficient is reduced by 3% to compensate for elastic deformation. For scenarios where the robot's end-effector linear velocity is >1m / s, centrifugal force correction is introduced. The angle correction amount is positively correlated with the square of the angular velocity (e.g., if the angular velocity increases by 20%, the correction amount increases by 4%). Combining the end-effector pose error ΔP, the corrected stiffness coefficient K', and the temperature change ΔT, the DH parameter adjustment value is output through the historical error-parameter mapping model (trained based on the running data of the previous 3 months) to ensure that the kinematic model adapts to the entity state in real time.

[0201] In this embodiment of the invention, the error trend for the next 30 seconds can be predicted by an LSTM neural network before calibration, and three sets of correction schemes are pre-generated. After correction, three sets of typical verification trajectories (including straight lines, circular arcs and composite curves) are executed, requiring the average error of the trajectory to be ≤0.6×P0 and the maximum error to be ≤P0. Otherwise, the calibration process is repeated, thereby achieving high-precision mapping between the virtual model and the physical entity.

[0202] This invention proposes a modeling technique that integrates 3D scanning and dynamic correction of kinematic parameters. By controlling the geometric error at the millimeter level and optimizing the motion prediction at the sub-millimeter level, it solves the problem of insufficient accuracy of traditional twin models and achieves accurate mapping of the virtual model to the physical robot.

[0203] Reference Figure 4 This diagram illustrates the logic of dynamic task allocation and optimization provided in an embodiment of the present invention. The method for dynamic task allocation and optimization may specifically include the following steps:

[0204] Step 401: Organize the tasks to be assigned and collect real-time load data for each robot;

[0205] Step 402: Quantify the load intensity using a weighted formula;

[0206] Step 403: Start allocation when the load index is greater than or equal to the adjustment threshold; otherwise, maintain the current state.

[0207] Step 404: Use an LSTM model to predict load changes in the next 5 minutes to identify overload risks in advance;

[0208] Step 405: Automatically select the mode according to the scenario (layered alternation for large-scale operations, priority rotation for tagged tasks, and capability complementarity for over-limit tasks);

[0209] Step 406: Pre-allocate 20%-40% of potential tasks to idle robots (to verify that there are no conflicts in the trajectory), optimize the final solution through multiple objective functions (balance, duration, loss, energy consumption), and execute it.

[0210] In this embodiment of the invention, to achieve accurate assessment of robot load, a multi-parameter fusion load intensity index is constructed: L = w1 × workload percentage + w2 × motion complexity + w3 × load change rate. Wherein, workload percentage = current executed task volume / rated daily task volume (reflecting task saturation); motion complexity = number of trajectory inflection points × average curvature (quantifying motion difficulty); load change rate = load fluctuation amplitude per unit time (reflecting load stability); the weights w1, w2, and w3 can be dynamically adjusted according to the real-time scenario (e.g., w2 increases to 0.4 for precision operations, and w1 increases to 0.5 for batch handling), and satisfy w1 + w2 + w3 = 1.

[0211] When L ≥ the warning threshold (e.g., 0.7), the system can enhance the monitoring frequency of joint temperature and current (from 1Hz to 5Hz); when L ≥ the adjustment threshold (e.g., 0.85), the dynamic task allocation process is triggered immediately.

[0212] For multi-robot collaborative scenarios, this invention provides three dynamically switchable operating modes to ensure load balancing and maximized efficiency.

[0213] Layered Alternating Mode: The workspace is divided into a three-dimensional grid (e.g., 1m × 1m × 0.5m). Robot A is responsible for the work in odd-numbered grid areas, while Robot B is responsible for even-numbered grid areas. Switching intervals are ≤0.5s through real-time coordinate comparison, avoiding path conflicts. This mode is suitable for large-area flat operations (such as warehouse sorting) and can reduce the distance that robots cross paths.

[0214] Priority rotation mode: Tasks are categorized by "urgency (E) - complexity (C)" (E and C are both levels 1-5). Under high load, tasks are rotated among robots according to the rule "E1C1→E2C2→...". For example, when robot A has a load L=0.9, the task with E=3 and C=4 is rotated to robot B with a load L=0.6, ensuring that the ∑(E×C) deviation between the two robots is ≤10% to avoid overloading a single device.

[0215] Complementary Capability Mode: Based on the robot parameter matrix (maximum load, maximum speed, positioning accuracy, etc.), when the task parameters (e.g., required load of 20kg) exceed the current robot's capability (e.g., maximum load of 15kg), the system automatically assigns the task to a suitable robot (e.g., robot C with a maximum load of 30kg). The task transition path is planned according to "starting position + adaptive speed integral" to ensure path smoothness (speed change rate ≤ 0.5m / s). 2 ).

[0216] To avoid efficiency drops caused by sudden load increases, a proactive control strategy of "prediction-pre-allocation-optimization" is implemented. Load trend prediction: The current load curve is fitted using a sliding window algorithm (window size 10 minutes), and the load curve L(t) for the next 5 minutes is predicted by combining it with historical data from the same period. When L(t) is predicted to be greater than or equal to the adjustment threshold, 20%-40% of the non-urgent tasks of the current robot are pre-allocated to idle robots, with the spatial overlap rate of the pre-allocated trajectories ≤5% (verified by a collision detection algorithm).

[0217] This invention also includes a multi-objective optimization model. Specifically, the core optimization function is a comprehensive objective function F = 0.4 × load balance + 0.3 × total time + 0.2 × equipment loss + 0.1 × energy consumption. Here, load balance = 1 - robot load standard deviation / average load; equipment loss is positively correlated with the number of joint movements and load duration; energy consumption is calculated as the integral of motor power. In emergency scenarios (such as fault repair), the weight of total time is increased to 0.5, while in normal scenarios, the weight of load balance is kept at 0.4. The optimal allocation scheme is then solved using a particle swarm optimization algorithm.

[0218] The present invention designs an intelligent scheduling strategy based on real-time monitoring of the load of the work unit, which breaks through the limitations of fixed division of labor, automatically switches between work modes such as alternating execution and priority rotation, reduces the continuous work time of a single robot by more than 46.7%, and significantly improves the collaborative efficiency of multiple robots.

[0219] Reference Figure 5 The diagram illustrates a logic diagram for handling virtual-real linkage faults and providing teaching evaluation according to an embodiment of the present invention. The method for handling virtual-real linkage faults and providing teaching evaluation may specifically include the following steps:

[0220] Step 501: Select the fault type (e.g., encoder malfunction) and start the multiphysics simulation of the virtual twin;

[0221] Step 502: The operator's physical actions are mapped to the virtual model via sensors, and the virtual system generates an early warning synchronously.

[0222] Step 503: Monitor joint angular velocity, obstacle distance, and operating force in real time. If any of these exceed the limit, freeze the operation and prompt for correction.

[0223] Step 504: After the operation meets the standards, the system evaluates the operation from three aspects: smoothness of operation, scientific nature of decision-making, and effectiveness of results.

[0224] Step 505: Generate a comprehensive score and personalized improvement suggestions.

[0225] In this embodiment of the invention, to achieve accurate fault reproduction and early warning, a multi-physics field coupled fault simulation module is constructed in a virtual twin. When an encoder signal jump is detected, the virtual model synchronously simulates the multi-physics field change. Temperature field: T(t) = T0 + temperature rise coefficient × fault duration (the temperature rise coefficient is set according to the degree of current overload, such as 0.5℃ / s when the current exceeds the rated value by 20%); Vibration characteristics: the vibration frequency increases exponentially from the normal 10Hz as the fault intensifies (e.g., reaching 50Hz 5s after the fault); Current waveform: random interference signal is superimposed, and the interference intensity increases linearly with the fault duration (e.g., the interference amplitude reaches 15% of the rated current after 10s).

[0226] The embodiments of the present invention also set a diffusion threshold (such as a temperature of 60°C or a vibration acceleration of 5 m / s²). 2 Once the threshold is exceeded, the probability of fault propagation is positively correlated with the current load (L value) and duration (the higher the load and the longer the duration, the higher the probability of propagation is every 10 seconds). The physical damage quantity D is quantified by the time integral value of temperature and vibration, providing a basis for assessing the consequences of the fault.

[0227] This invention also constructs a closed-loop interactive system of "physical operation - virtual feedback - physical adjustment" to support safe teaching and remote operation. Specifically, the operator wears motion capture equipment (such as an inertial sensor), and their hand posture (angle, displacement) is wirelessly transmitted and mapped to the virtual robot. The mapping error is ≤±5° ​​(the sensor delay is corrected through a calibration algorithm), ensuring consistency between virtual operation and actual operation.

[0228] When the virtual robot enters a dangerous critical zone (e.g., ≤10cm from an obstacle), the system generates a red warning trajectory and triggers multiple warnings on the physical end: the buzzer frequency increases as the distance to the danger zone decreases (e.g., the frequency reaches 2kHz at 5cm); voice prompts provide specific adjustment angles (e.g., "Please adjust 3° to the left"), with the adjustment amount calculated in real time based on the obstacle avoidance algorithm; and the virtual scene displays simulated collision consequences (e.g., component deformation animation) to enhance the operator's risk awareness.

[0229] To ensure operational safety and teaching effectiveness, this invention establishes a three-dimensional evaluation system that monitors three key safety indicators in real time: joint angular velocity ≤ safety threshold (e.g., 180° / s); distance to obstacles ≥ 5cm (detected in real time by lidar); and end effector force ≤ 50N (feedback from force sensors). If any indicator exceeds the limit, the virtual operation is frozen, a correction plan is generated (e.g., deceleration command, path deviation suggestion), and the system is unlocked after adjustments are made to meet the requirements.

[0230] This invention provides a quantitative scoring system for operators' workflow: Operational fluency S1: a reverse scoring system based on the percentage of operations exceeding the standard time limit (100 points for no timeout, 80 points for 20% timeout); Decision-making scientificity S2: a comparison of the operational steps with the standard procedure (100 points for perfect alignment, 20 points deducted per key step deviation); Result effectiveness S3: a scoring system combining the percentage of virtual damage (Dactual / Dmaximum) and the failure recovery time compliance rate (100 points for damage ≤10% and compliance with time requirement). The comprehensive score S = S1 × 0.3 + S2 × 0.4 + S3 × 0.3, and personalized improvement plans are output based on the score (e.g., reinforcing standard procedure instruction for those with low S2 scores).

[0231] This invention constructs a real-time synchronous response mechanism between virtual and physical systems under fault conditions. It combines collision-free operation simulation verification and multi-dimensional (response time, compliance, cost) quantitative evaluation to solve the problem of traditional training lacking safety verification and objective evaluation, and forms a standardized teaching evaluation scheme.

[0232] Reference Figure 6 This diagram illustrates a logical diagram of a multi-technology fusion system architecture provided by an embodiment of the present invention. The method for multi-technology fusion may specifically include the following steps:

[0233] Step 601: Collect physical robot data, environmental parameters, task instructions, and user operation information;

[0234] Step 602: Classify and process the input data, prioritize the transmission of key signals such as faults, compress point cloud data to retain core features, and associate geometric and physical parameters;

[0235] Step 603, Twin Modeling Module: Responsible for high-precision modeling and dynamic calibration; Task Allocation Module: Implements load quantization, pattern matching, and optimized scheduling; Fault Handling Module: Completes fault simulation, virtual-real interaction, and security verification;

[0236] Step 604: Quickly switch between industrial and teaching scenarios through modular components, predict the remaining lifespan of equipment based on load and temperature data, and achieve bidirectional complementarity between industrial and teaching data (industrial failure cases are added to the teaching database, and teaching errors are optimized for industrial early warning).

[0237] In this embodiment of the invention, an intelligent data fusion engine is designed to address the heterogeneity and redundancy issues of multi-source data. Specifically, data is prioritized according to three dimensions: timeliness, accuracy, and relevance. Fault signals (such as emergency stop signals) have the highest priority (real-time requirement ≤10ms), followed by status signals such as ambient temperature, and historical log data has the lowest priority.

[0238] Point cloud data is processed using a feature-preserving compression algorithm to ensure that ≥90% of key feature points (such as edges and holes) are retained, with a compression ratio of 1:5; sensor data is smoothed by Kalman filtering to eliminate high-frequency noise (the fluctuation amplitude of the filtered data is ≤5% of the original data).

[0239] To improve the system's versatility and maintenance efficiency, this embodiment of the invention incorporates modular expansion and health management functions. Specifically, a modular component library (including scenario templates for welding, assembly, and handling) is established. When changing scenarios, simply calling the corresponding template allows the system to automatically adjust parameters (such as scanning accuracy and load threshold). For example, when switching to a welding scenario, the temperature sensitivity threshold is automatically lowered to 3°C, and the load evaluation weight w2 (motion complexity) is increased to 0.4.

[0240] This invention can also quantify the degree of equipment aging by using the integral loss of remaining life L_remain = rated life - load × temperature (the integral period is the daily operating time). When L_remain is lower than a threshold (e.g., 20% of the rated life), the system automatically generates a maintenance work order, which includes the parts to be replaced (e.g., spherical bearings), the recommended maintenance time, and the spare parts model, thus achieving predictive maintenance.

[0241] This invention innovatively integrates three major technical modules: high-precision modeling, dynamic scheduling, and virtual-real interaction, achieving full-process optimization from model building and task execution to teaching verification, and providing a systematic solution for robot digital twins in the industrial and educational fields.

[0242] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0243] Reference Figure 7 The diagram illustrates a structural block diagram of a robot task allocation device based on digital twins provided by an embodiment of the present invention, which may specifically include the following modules:

[0244] The virtual twin model generation module 701 is used to scan multiple robots and generate virtual twin models corresponding to the multiple robots based on the scanning results.

[0245] The load assessment information determination module 702 is used to obtain the currently assigned tasks of the plurality of robots and determine the load assessment information of the plurality of robots executing the currently assigned tasks.

[0246] The robot to be assigned module 703 is used to determine the target robot to be assigned a task among the multiple robots based on the load assessment information of the multiple robots performing the currently assigned tasks.

[0247] The simulation allocation result determination module 704 is used to determine the collaborative operation scenario, and to perform task allocation simulation according to the virtual twin models corresponding to the multiple robots and the collaborative operation scenario to obtain the simulation allocation result;

[0248] The simulated target task allocation module 705 is used to allocate tasks to the target robot to be assigned tasks based on the simulation allocation results.

[0249] In this embodiment of the invention, the device further includes:

[0250] An abnormal robot detection module is used to identify the abnormal robot among the multiple robots when an abnormality is detected.

[0251] The virtual repair operation evaluation module is used to detect and evaluate the virtual repair operations performed by the operator on the abnormal robot.

[0252] In this embodiment of the invention, the load assessment information determination module 702 includes:

[0253] The load assessment information acquisition submodule is used to determine the task saturation, motion complexity, and load change rate of each robot.

[0254] The load assessment information determination submodule is used to determine the load assessment information of each robot for performing the currently assigned task based on the task saturation, motion complexity and / or load change rate of each robot.

[0255] In this embodiment of the invention, the collaborative work scenario includes: a layered alternating scenario; the simulation allocation result determination module 704 includes:

[0256] The simulated work area determination submodule is used to determine the virtual position coordinates of each robot when the collaborative work scenario of the multiple robots is a layered and alternating scenario, and to perform path movement simulation on the virtual twin model corresponding to the multiple robots based on the virtual position coordinates of each robot to obtain the simulated work area of ​​the target robot of the task to be assigned.

[0257] The simulated target task allocation module 705 includes:

[0258] The first simulated target task allocation submodule is used to allocate tasks to the target robot based on the simulated working area of ​​the target robot to be assigned the task.

[0259] In this embodiment of the invention, the collaborative work scenario includes: a priority rotation scenario; the simulated allocation result determination module 704 includes:

[0260] The simulation rotation result determination submodule is used to obtain the task level of the currently assigned tasks of each robot when the collaborative operation scenario of the multiple robots is a priority rotation scenario; determine the task deviation between each robot based on the task level of the currently assigned tasks of each robot and the load assessment information of each robot; and perform task rotation simulation on the virtual twin model corresponding to each robot based on the task deviation between each robot to obtain the simulation rotation result of the target robot of the task to be assigned.

[0261] The simulated target task allocation module 705 includes:

[0262] The second simulated target task allocation submodule is used to allocate tasks to the target robots of the tasks to be assigned based on the simulated rotation results of the target robots of the tasks to be assigned.

[0263] In this embodiment of the invention, the collaborative operation scenario includes: a capability complementarity scenario; the simulation allocation result determination module 704 includes:

[0264] The simulation transition result determination submodule is used to obtain the load capacity of each robot and the load parameters of the currently assigned tasks of each robot when the collaborative operation scenario of the multiple robots is a complementary capability scenario; based on the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, the module performs task transition simulation on the virtual twin models corresponding to the multiple robots to obtain the simulation transition result of the target robot of the task to be assigned.

[0265] The simulated target task allocation module 705 includes:

[0266] The third simulated target task allocation submodule is used to allocate tasks to the target robots based on the simulation transition results of the task to be assigned.

[0267] In this embodiment of the invention, the robot to be assigned determination module 703 includes:

[0268] The robot to be assigned submodule is used to select robots whose load assessment information is greater than or equal to a preset load threshold from among the multiple robots as target robots for the task to be assigned.

[0269] In this embodiment of the invention, the device further includes:

[0270] The equipment information determination module is used to obtain the load balance, total task duration, equipment wear and tear, and equipment energy consumption of the multiple robots; the load balance is determined based on the load standard deviation and average load of the multiple robots; the equipment wear and tear is determined based on the number of joint movements and load assessment information of the robots; and the equipment energy consumption is determined based on the motor power of the robots.

[0271] The task allocation optimization module is used to optimize the tasks allocated to the multiple robots based on the load balance of the multiple robots, the total task duration, equipment wear and tear, and equipment energy consumption.

[0272] In this embodiment of the invention, the virtual twin model generation module 701 includes:

[0273] A 3D data generation submodule is used to perform 3D scanning on the structure of the multiple robots to obtain 3D data of the multiple robots respectively;

[0274] The three-dimensional data preprocessing submodule is used to preprocess the three-dimensional data of the multiple robots to obtain the initial virtual twin models corresponding to the multiple robots respectively;

[0275] The state parameter acquisition submodule is used to acquire the state parameters of the multiple robots respectively;

[0276] The virtual twin model generation submodule is used to correct the initial virtual twin models corresponding to the multiple robots when the state parameters of the multiple robots meet the preset conditions, so as to obtain the virtual twin models corresponding to the multiple robots respectively.

[0277] In this embodiment of the invention, the state parameters of the plurality of robots include the three-dimensional pose of the robot's end effector, the joint angle of the robot, the operating current of the robot, and the ambient temperature.

[0278] The preset conditions include: the robot's end-effector pose error is greater than a preset pose error; and / or, the robot's operating current exceeds a preset current range; and / or, the ambient temperature changes more than a preset temperature threshold within a preset time period.

[0279] The virtual twin model generation submodule includes:

[0280] The joint stiffness coefficient update unit is used to obtain the current fluctuation influence coefficient of the multiple robots, and update the joint stiffness coefficient of the initial virtual twin model corresponding to the multiple robots according to the current fluctuation influence coefficient of the multiple robots.

[0281] The joint angle update unit is used to obtain the joint angular velocities of the multiple robots and update the joint angles of the initial virtual twin models corresponding to the multiple robots based on the joint angular velocities of the multiple robots.

[0282] The virtual twin model generation unit is used to acquire the change in ambient temperature and determine the virtual twin model corresponding to the multiple robots based on the end-effector pose error of the multiple robots, the joint stiffness coefficient after correction of the initial virtual twin model corresponding to the multiple robots, the joint angle after correction of the initial virtual twin model corresponding to the multiple robots, and the change in ambient temperature.

[0283] In this embodiment of the invention, the virtual repair operation evaluation module includes:

[0284] The virtual repair operation determination submodule is used to determine the smoothness of the operation, the consistency of the operation steps, and the validity of the operation results of the virtual repair operation performed by the operator on the abnormal robot; wherein, the smoothness of the operation is determined based on the operation duration of the virtual repair operation and the preset duration, the consistency of the operation steps is determined based on the consistency of the operation steps of the virtual repair operation with the preset process, and the validity of the operation results is determined based on the virtual damage ratio and the failure recovery time achievement rate of the virtual repair operation;

[0285] The virtual repair operation evaluation submodule is used to evaluate the operator's operation process based on the smoothness of the operation, the consistency of the operation steps, and the validity of the operation results.

[0286] In this embodiment of the invention, the device further includes:

[0287] An abnormal robot information acquisition module is used to acquire the joint angular velocity of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot.

[0288] The virtual repair operation stop module is used to stop the operator's virtual repair operation when the joint angular velocity of the abnormal robot is greater than or equal to a preset speed threshold, and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold, and / or the end effector force of the abnormal robot is greater than or equal to a preset effector force threshold.

[0289] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0290] This invention also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described robot task allocation method based on digital twin.

[0291] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described robot task allocation method based on digital twins.

[0292] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0293] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products embodied on one or more machine-readable media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0294] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0295] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0296] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0297] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0298] Finally, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device 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 terminal device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0299] The above provides a detailed description of the robot task allocation method and apparatus based on digital twins provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A robot task allocation method based on digital twins, characterized in that, The method includes: Multiple robots are scanned, and virtual twin models corresponding to the multiple robots are generated based on the scan results; Obtain the currently assigned tasks of the multiple robots, and determine the load assessment information of the multiple robots executing the currently assigned tasks; Based on the load assessment information of the multiple robots performing the currently assigned tasks, determine the target robot among the multiple robots to be assigned tasks; Determine the collaborative operation scenario, and based on the virtual twin models corresponding to the multiple robots, simulate task allocation according to the collaborative operation scenario to obtain the simulation allocation result; Based on the simulation allocation results, the target robot to be assigned the task is assigned a task.

2. The method according to claim 1, characterized in that, The method further includes: When an anomaly is detected among the plurality of robots, the abnormal robot among the plurality of robots is identified; The operator performs virtual repair operations on the malfunctioning robot and evaluates these virtual repair operations.

3. The method according to claim 1, characterized in that, The determination of the load assessment information for the multiple robots to perform the currently assigned tasks includes: Determine the task saturation, motion complexity, and load change rate for each robot; Based on the task saturation, motion complexity, and / or load change rate of each robot, determine the load assessment information for each robot to perform the currently assigned task.

4. The method according to claim 1, characterized in that, The collaborative operation scenario includes: a layered alternating scenario; the step of simulating task allocation according to the virtual twin models corresponding to the multiple robots and obtaining the simulation allocation result includes: In the case where the collaborative operation scenario of the multiple robots is a layered and alternating scenario, the virtual position coordinates of each robot are determined, and the path movement simulation of the virtual twin model corresponding to the multiple robots is performed based on the virtual position coordinates of each robot to obtain the simulated operation area of ​​the target robot of the task to be assigned. The step of assigning tasks to the target robot based on the simulation assignment results includes: Tasks are assigned to the target robot based on the simulated working area of ​​the target robot to be assigned the task.

5. The method according to claim 1, characterized in that, The collaborative operation scenario includes: a priority rotation scenario; the step of simulating task allocation according to the collaborative operation scenario based on the virtual twin models corresponding to the multiple robots, and obtaining the simulation allocation result, includes: In the case where the collaborative operation scenario of the multiple robots is a priority rotation scenario, the task level of the currently assigned tasks of each robot is obtained; based on the task level of the currently assigned tasks of each robot and the load assessment information of each robot, the task deviation between each robot is determined; based on the task deviation between each robot, the virtual twin model corresponding to each robot is used to simulate task rotation, and the simulation rotation result of the target robot of the task to be assigned is obtained. The step of assigning tasks to the target robot based on the simulation assignment results includes: Based on the simulation rotation results of the target robots for the tasks to be assigned, tasks are assigned to the target robots for the tasks to be assigned.

6. The method according to claim 1, characterized in that, The collaborative operation scenario includes: a capability complementarity scenario; the step of simulating task allocation according to the collaborative operation scenario based on the virtual twin models corresponding to the multiple robots, and obtaining the simulation allocation result, includes: In the case where the collaborative operation scenario of the multiple robots is a scenario of complementary capabilities, the load capacity of each robot and the load parameters of the currently assigned tasks of each robot are obtained; based on the load capacity of each robot and the load parameters of the currently assigned tasks of each robot, a task transition simulation is performed on the virtual twin model corresponding to the multiple robots to obtain the simulation transition result of the target robot of the task to be assigned. The step of assigning tasks to the target robot based on the simulation assignment results includes: The simulation transition results of the target robot to be assigned the task are used to assign the task to the target robot.

7. The method according to claim 1, characterized in that, The step of determining the target robot to be assigned a task among the multiple robots based on the load assessment information of the multiple robots performing the currently assigned tasks includes: Robots whose load assessment information is greater than or equal to a preset load threshold are selected as target robots for task assignment.

8. The method according to claim 1, characterized in that, The method further includes: The load balance, total task duration, equipment wear and tear, and equipment energy consumption of the multiple robots are obtained; the load balance is determined based on the load standard deviation and average load of the multiple robots; the equipment wear and tear is determined based on the number of joint movements and load assessment information of the robots; and the equipment energy consumption is determined based on the motor power of the robots. The tasks assigned to the multiple robots are optimized based on the load balancing, total task duration, equipment wear and tear, and equipment energy consumption.

9. The method according to claim 1, characterized in that, The step of scanning multiple robots and generating virtual twin models corresponding to the multiple robots based on the scanning results includes: The structures of the multiple robots are subjected to three-dimensional scanning to obtain the three-dimensional data of the multiple robots respectively; The 3D data of the multiple robots are preprocessed to obtain the initial virtual twin models corresponding to the multiple robots; Obtain the state parameters of the multiple robots respectively; When the state parameters of the multiple robots meet the preset conditions, the initial virtual twin models corresponding to the multiple robots are corrected to obtain the virtual twin models corresponding to the multiple robots respectively.

10. The method according to claim 9, characterized in that, The state parameters of the multiple robots include the three-dimensional pose of the robot's end effector, the joint angles of the robot, the operating current of the robot, and the ambient temperature. The preset conditions include: the robot's end-effector pose error is greater than a preset pose error; and / or, the robot's operating current exceeds a preset current range; and / or, the ambient temperature changes more than a preset temperature threshold within a preset time period. The step of correcting the initial virtual twin models corresponding to the plurality of robots to obtain virtual twin models corresponding to the plurality of robots includes: Obtain the current fluctuation influence coefficient of the multiple robots, and update the joint stiffness coefficient of the initial virtual twin model corresponding to the multiple robots based on the current fluctuation influence coefficient of the multiple robots. The joint angular velocities of the multiple robots are obtained, and the joint angles of the initial virtual twin models corresponding to the multiple robots are updated based on the joint angular velocities of the multiple robots. The change in ambient temperature is obtained, and the virtual twin model corresponding to the multiple robots is determined based on the end-effector pose error of the multiple robots, the joint stiffness coefficient after correction of the initial virtual twin model corresponding to the multiple robots, the joint angle after correction of the initial virtual twin model corresponding to the multiple robots, and the change in ambient temperature.

11. The method according to claim 2, characterized in that, The detection operator performs virtual repair operations on the abnormal robot and evaluates the virtual repair operations, including: The smoothness of the operation, the consistency of the operation steps, and the validity of the operation results of the virtual repair operation performed by the operator on the abnormal robot are determined; wherein, the smoothness of the operation is determined based on the operation time of the virtual repair operation and the preset time, the consistency of the operation steps is determined based on the consistency of the operation steps of the virtual repair operation with the preset process, and the validity of the operation results is determined based on the virtual damage ratio and the failure recovery time achievement rate of the virtual repair operation. The operator's operation process is evaluated based on the smoothness of the virtual repair operation, the consistency of the operation steps, and the validity of the operation results.

12. The method according to claim 2, characterized in that, The method further includes: The joint angular velocities of the abnormal robot, the distance between the abnormal robot and the obstacle, and the end effector force of the abnormal robot are obtained. The virtual repair operation performed by the operator shall be stopped if the joint angular velocity of the abnormal robot is greater than or equal to a preset velocity threshold, and / or the distance between the abnormal robot and the obstacle is less than or equal to a preset distance threshold, and / or the end effector force of the abnormal robot is greater than or equal to a preset effector force threshold.

13. A robot task allocation device based on digital twin, characterized in that, The device includes: The virtual twin model generation module is used to scan multiple robots and generate virtual twin models corresponding to the multiple robots based on the scanning results. The load assessment information determination module is used to obtain the currently assigned tasks of the multiple robots and determine the load assessment information of the multiple robots executing the currently assigned tasks. The robot to be assigned module is used to determine the target robot among the multiple robots to be assigned the task based on the load assessment information of the multiple robots performing the currently assigned task; The simulation allocation result determination module is used to determine the collaborative operation scenario, and to perform task allocation simulation according to the virtual twin models corresponding to the multiple robots and the collaborative operation scenario to obtain the simulation allocation result; The simulated target task allocation module is used to allocate tasks to the target robot to be assigned tasks based on the simulated allocation results.

14. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of a robot task allocation method based on a digital twin as described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of a robot task allocation method based on digital twins as described in any one of claims 1-12.

Images