An intelligent sorting management system for industrial robots

Abstract

The application relates to the technical field of intelligent sorting, in particular to an intelligent sorting management system for an industrial robot, which comprises a digital twin module, a data acquisition module, a visual recognition module, a sorting execution module, a perception module, an exception determination module and an adjustment module. The application completes the whole-process simulation of the grabbing and placing path in advance through the digital twin module, combines the real-time monitoring of multi-dimensional information through the data acquisition module, forms high-precision operation state perception, the visual recognition module ensures the accurate recognition of material categories, goods grid wear states and space coordinates, the perception module realizes dynamic monitoring of operation risks through comprehensive evaluation of three-dimensional states of stability, smoothness and compliance, the exception determination module determines exception types based on multi-parameter cross calculation, avoids single-index misjudgment and realizes high-precision and strong-robustness intelligent sorting operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent sorting, in particular to an intelligent sorting management system for industrial robots. BACKGROUND

[0002] With the development of intelligent manufacturing and industrial automation, the precision, efficiency and flexibility requirements of modern warehousing and logistics for sorting systems are becoming higher and higher. Traditional mechanical sorting systems are difficult to cope with the uncertainty problems caused by the variety of material forms, complex grasping posture and changes in working environment due to the dependence on fixed paths and manual parameter setting, resulting in grasping failure, material damage and overall sorting efficiency decline, so an intelligent solution that can realize real-time perception, dynamic adjustment and optimization of sorting strategy is urgently needed.

[0003] Chinese patent application publication No. CN112705488A discloses an intelligent sorting management system, which comprises a rack, a feeding hopper arranged on the rack, a lifting device for lifting materials to a high place, a code scanning mechanism arranged below the top of the lifting device, a conveying mechanism arranged below the code scanning mechanism, a material stirring mechanism arranged above the conveying mechanism, and a material receiving device arranged on one side of the conveying mechanism for receiving the materials stirred down by the material stirring mechanism.

[0004] Therefore, the intelligent sorting management system has the following problems: the micro-mechanics and localization of the system lack perception, leading to unstable grasping; the path planning and adjustment of the system have no simulation verification and are difficult to safely adapt; the abnormality determination of the system is linear and has no closed-loop adaptation, leading to frequent manual intervention. SUMMARY

[0005] Therefore, the present application provides an intelligent sorting management system for industrial robots to overcome the problems of grasping failure, material damage and low overall sorting efficiency caused by unstable grasping force, material micro-sliding and deviation of placement posture in the grasping process in the prior art through multi-parameter real-time perception and closed-loop adaptive adjustment.

[0006] To achieve the above-mentioned purpose, the present application provides an intelligent sorting management system for industrial robots, comprising:

[0007] A digital twin module is used to perform path simulation in a virtual environment based on a preset three-dimensional robot model according to the real-time received grasping and placing task list before physical execution, to generate control instructions;

[0008] A data acquisition module is used to acquire in real time the industrial camera images when the materials placed at a preset placement angle arrive at each sorting station, the micro-contact pressure deviation of the gripper of each industrial robot during the grasping process, the change rate of the grasping path curvature, and the micro-sliding deviation of the materials;

[0009] The visual recognition module is used to identify the physical spatial coordinates, gripping posture, and edge wear area of ​​the material based on the image from the industrial camera.

[0010] The sorting execution module is used to drive the industrial robot to perform sorting actions according to the control instructions and the physical space coordinates, so as to place the material in the target grid of the pallet.

[0011] The sensing module is used to sense the stability state based on the micro-contact pressure deviation, the micro-slip deviation and the edge wear area, as well as to sense the motion smoothness state based on the rate of change of curvature of the grasping path, and to sense the compliance state based on the grasping posture.

[0012] An anomaly determination module is used to determine several anomaly types based on the stability state, the motion smoothness state, and the compliance state.

[0013] An adjustment module is used to adjust the preset gripper torque and / or preset grasping speed and / or preset placement angle of the industrial robot according to the anomaly type and a preset adjustment mapping table.

[0014] Furthermore, the digital twin module includes:

[0015] The extraction unit is used to parse and extract the grasping pose, placement pose, and operation sequence of the material according to the grasping and placement task list, and obtain the parsing result;

[0016] The task compilation unit, which is connected to the extraction unit, is used to map the parsing results to the robot arm base coordinate system and the workstation coordinate system, and to generate a constrained process step sequence in combination with preset process constraints.

[0017] The simulation unit, which is connected to the task compilation unit, is used to perform path planning and motion simulation on the constrained process sequence based on a preset three-dimensional robotic arm model, so as to generate control instructions for driving the sorting execution.

[0018] Furthermore, the simulation unit includes:

[0019] The parameter determination subunit is used to perform inverse kinematics solution and time parameterized trajectory planning on the constrained work step sequence based on the preset three-dimensional robotic arm model in a virtual environment, so as to obtain the expected position, velocity and acceleration of each joint and determine the motion parameters.

[0020] The trajectory timing generation subunit is connected to the parameter determination subunit to generate a complete joint space running trajectory and end effector action timing based on the motion parameters.

[0021] A constraint subunit, which is connected to the trajectory timing generation subunit, is used to perform full-process obstacle avoidance and collision detection according to the preset work station layout and preset obstacle modeling information, so as to constrain the joint space running trajectory and obtain the constrained running trajectory.

[0022] The instruction generation subunit is connected to the trajectory timing generation subunit and the constraint subunit respectively, and is used to package the constrained running trajectory and the end effector action timing according to the controller communication protocol to generate the control instruction.

[0023] Furthermore, the sensing module includes:

[0024] A stable sensing unit is used to calculate a stability index based on the micro-contact pressure deviation, the micro-slip deviation, and the edge wear area within a preset sensing time, and to sense the stability state based on the stability index.

[0025] A stability sensing unit is used to calculate a stability index based on the rate of change of curvature of the grasping path within the same preset sensing time, and to sense the stability state based on the stability index.

[0026] The compliance perception unit is used to calculate a compliance index based on the posture and a preset posture library within the same preset perception time period, and to perceive the compliance status based on the compliance index.

[0027] Furthermore, the stable sensing unit includes:

[0028] The stabilization subunit is used to perform maximum-minimum normalization processing on the micro-contact pressure deviation at the end of the preset sensing time based on all the micro-contact pressure deviations within the preset sensing time to obtain a pressure deviation normalization value; and to perform maximum-minimum normalization processing on the micro-slip deviation at the end of the preset sensing time based on all the micro-slip deviations within the preset sensing time to obtain a slip deviation normalization value; and to perform maximum-minimum normalization processing on the edge wear area at the end of the preset sensing time based on all the edge wear areas within the preset sensing time to obtain a wear area normalization value.

[0029] The stability index calculation subunit, which is connected to the stability normalization subunit, is used to perform a weighted summation calculation on the pressure deviation normalization value, the preset pressure deviation weight, the sliding deviation normalization value, the preset sliding deviation weight, the wear area normalization value, and the preset wear area weight to obtain the stability index.

[0030] A stability sensing subunit, which is connected to the stability index calculation subunit, is used to determine that the stability state is sensed when the stability index is greater than a preset stability index threshold.

[0031] Furthermore, the stable sensing unit includes:

[0032] The steady-state normalization subunit is used to calculate the standard deviation of the rate of change of curvature of all the grasping paths from the initial time to each time within the same preset perception time, to obtain several curvature change fluctuation values, and to perform maximum-minimum normalization processing on each curvature change fluctuation value to obtain several fluctuation normalization values.

[0033] A stability index calculation subunit, which is connected to the stability normalization subunit, is used to calculate the average value of all the fluctuation normalization values ​​to obtain the stability index.

[0034] A stability sensing subunit, which is connected to the stability index calculation subunit, is used to determine that the stability state is sensed when the stability index is greater than a preset stability index threshold.

[0035] Furthermore, the compliance awareness unit includes:

[0036] The matching degree calculation subunit is used to compare the grasping posture at each moment within the same preset perception time with the reference posture in the preset posture library to obtain several matching degrees.

[0037] A compliance index calculation subunit, which is connected to the matching degree calculation subunit, is used to calculate the average value of all the matching degrees to obtain the compliance index;

[0038] A compliance perception subunit, which is connected to the compliance index calculation subunit, is used to determine the perceived compliance status when the compliance index is greater than a preset compliance index threshold.

[0039] Furthermore, the anomaly determination module includes:

[0040] The contribution value calculation unit is used to calculate a stable contribution value based on a stability index, a stable contribution value based on a stability index, and a compliance contribution value based on a compliance index when the stability state, the stability state, and the compliance state are perceived.

[0041] An anomaly index calculation unit is used to calculate an anomaly index based on the stable contribution value, the steady contribution value, and the compliant contribution value.

[0042] The type determination unit is used to determine several of the anomaly types based on the anomaly index, the stable contribution value, the steady contribution value, and the compliance contribution value.

[0043] Furthermore, the type determination unit is used to determine that the anomaly type is the first type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is less than a preset stable contribution threshold, the steady contribution value is greater than a preset steady contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold.

[0044] The type determination unit is further configured to determine the anomaly type as the second type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is greater than or equal to a preset stable contribution threshold, the steady contribution value is less than or equal to a preset steady contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold.

[0045] The type determination unit is further configured to determine that the anomaly type is a third type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is less than a preset stable contribution threshold, the steady contribution value is less than or equal to a preset steady contribution threshold, and the compliance contribution value is less than or equal to a preset compliance contribution threshold.

[0046] Furthermore, the adjustment module includes:

[0047] The mapping unit is used to map the first type, the second type and the third type to the preset adjustment mapping table respectively, so as to obtain the first adjustment instruction corresponding to the first type, the second adjustment instruction corresponding to the second type and the third adjustment instruction corresponding to the third type;

[0048] An adjustment unit, connected to the mapping unit, is used to adjust the preset gripper torque according to a preset torque adjustment coefficient based on the first adjustment command, and to adjust the preset gripping speed according to a preset speed adjustment coefficient based on the second adjustment command.

[0049] The adjustment unit is further configured to, based on the third adjustment instruction, adjust the preset gripper torque according to the preset stable contribution threshold and the stable contribution value, adjust the preset gripping speed according to the preset stable contribution threshold and the stable contribution value, and adjust the preset placement angle according to the compliance contribution value and the preset compliance contribution threshold.

[0050] Compared with existing technologies, the advantages of this invention are as follows: A digital twin module pre-simulates the entire process of gripping and placing the object, and a data acquisition module monitors in real time multi-dimensional information such as gripper pressure deviation, gripping path curvature change rate, and material micro-slip deviation, forming a high-precision operational status perception. A visual recognition module ensures accurate identification of material categories, compartment wear status, and spatial coordinates, guaranteeing the precision of path planning and execution. A perception module achieves dynamic monitoring of operational risks through a comprehensive evaluation of stability, smoothness, and compliance in three dimensions. Anomaly determination module determines anomaly types based on multi-parameter cross-calculation, avoiding misjudgment based on a single indicator. An adjustment module performs multi-parameter linkage adjustment of torque, speed, and placement angle based on the anomaly type and real-time parameter deviation, enabling force control and motion control to dynamically match material characteristics and the operating environment. This significantly improves sorting efficiency and gripping stability, reduces sorting failure rates caused by overload, misalignment, or slippage, and achieves high-precision, robust intelligent sorting operations.

[0051] Furthermore, by extracting the material gripping pose, placement pose, and operation sequence from the gripping and placement task list, and mapping the analysis results to the robotic arm base coordinate system and the workstation coordinate system, and simultaneously generating a constrained work step sequence based on preset process constraints, path planning and motion simulation can be performed in a virtual environment based on a preset 3D robotic arm model, thereby generating precise control commands. This process achieves dynamic matching and optimization between gripping torque, gripping speed, and placement angle by coordinating the material spatial position, gripping posture, path curvature, and robotic arm motion parameters, enabling the robotic arm to grip stably, move smoothly, and place accurately during physical execution, thereby improving overall sorting efficiency and operational reliability.

[0052] Furthermore, by performing inverse kinematics solving and time-parameterized trajectory planning on the constrained work step sequence, the expected position, velocity, and acceleration of each joint at different time points are accurately calculated, thereby forming a complete joint spatial running trajectory and end effector action sequence. Combined with the preset workstation layout and obstacle modeling information, the system can constrain the trajectory based on obstacle avoidance and collision detection throughout the process, ensuring that the robotic arm maintains stable movement in actual operation and can efficiently and accurately complete the grasping and placement tasks. At the same time, by packaging the trajectory and action sequence according to the controller protocol to generate control instructions, a high consistency between the virtual environment and physical execution is achieved, ensuring the safety and reliability of the sorting action.

[0053] Furthermore, by comprehensively analyzing micro-contact pressure deviation, micro-slippage deviation, and wear area at the edge of the gripper, the stability of the material during the gripping process is quantified. The smoothness of the robotic arm's movement is assessed by the rate of change of curvature along the gripping path. Simultaneously, a compliance index is calculated by comparing the gripping posture with a preset posture library, thus forming a multi-dimensional state perception of the gripping process. This multi-parameter joint analysis enables the system to accurately reflect the real-time performance of the robotic arm in terms of material force, motion trajectory, and gripping posture during actual operation. Through the mutual mapping and weighting of stability, smoothness, and compliance indices, it can effectively guide the subsequent optimization and adjustment of gripper torque, gripping speed, and placement angle, achieving dynamic monitoring and intelligent control of the sorting process, improving the gripping success rate, and reducing the risk of material slippage and damage.

[0054] Furthermore, by normalizing the micro-contact pressure deviation, micro-slip deviation, and wear area at the edge of the storage compartment within a preset sensing time, various raw data are mapped to a unified dimensional range to eliminate the influence of differences in the scale of different parameters. Subsequently, a stability index is calculated by weighted summation based on the preset weights of each parameter. This index comprehensively reflects the mechanical response, material sliding trend, and wear of the contact surface during the robotic arm's grasping process. Thus, when the stability index exceeds a preset threshold, the stability state can be accurately determined, enabling dynamic perception of the mechanical properties of the grasping action and the state of the material. This ensures that the control accuracy and safety of the industrial robot in the sorting process are effectively improved.

[0055] Furthermore, by normalizing the standard deviation of the rate of change of curvature of the grasping path within a preset sensing time, the curvature change fluctuation value is calculated, and the average value is further calculated to obtain a stability index. This stability index reflects the smoothness of the robotic arm's movement during the grasping process. By comparing the stability index with a preset stability index threshold, the system can accurately determine whether the movement is stable, thereby guiding the optimization of grasping speed and trajectory. This reduces vibration and deviation when the robotic arm performs sorting actions, ensuring smooth sliding of materials during grasping and placement, and improving overall sorting efficiency and reliability.

[0056] Furthermore, by comparing the gripping posture at each moment during the gripping process with reference postures in a preset posture library, the matching degree is obtained, and the average of all matching degrees is calculated to form a compliance index, which reflects the accuracy of the gripping action in terms of spatial position and posture. When the compliance index is higher than a preset threshold, the system determines that the gripping posture is compliant, thereby ensuring the accuracy of the relative position and angle between the gripper and the material, reducing tilting or rotation errors. At the same time, in conjunction with stability and smoothness parameters, it helps to achieve safe gripping, smooth transportation and accurate placement of materials, improving overall sorting efficiency and reliability.

[0057] Furthermore, by converting the stability index, smoothness index, and compliance index into corresponding contribution values, and then comprehensively calculating the anomaly index based on preset weights, the potential anomalies in the sorting process can be quantitatively determined. The stability contribution value, smoothness contribution value, and compliance contribution value reflect the force control accuracy, motion smoothness, and compliance degree of material gripping, respectively. The anomaly index, on the other hand, integrates the weighted influence of each contribution value, which can clearly distinguish different types of abnormal states and provide a scientific basis for subsequent adjustments to the gripper torque, gripping speed, and placement angle. This enables dynamic optimization and reliability improvement of robot actions in complex sorting environments.

[0058] Furthermore, by converting the stability index, smoothness index, and compliance index into corresponding contribution values, and then comprehensively calculating the anomaly index based on preset weights, the potential anomalies in the sorting process can be quantitatively determined. The stability contribution value, smoothness contribution value, and compliance contribution value reflect the force control precision, motion smoothness, and compliance degree of material gripping, respectively. The anomaly index, by comprehensively considering the weighted influence of each contribution value, can clearly distinguish different types of abnormal states and provide a scientific basis for subsequent adjustments to gripper torque, gripping speed, and placement angle. This enables dynamic optimization and reliability improvement of robot actions in complex sorting environments.

[0059] Furthermore, by quantifying and integrating mechanical and surface parameters such as micro-contact pressure deviation, material micro-slip deviation, and wear area of ​​the gripper edge with kinematic parameters such as the rate of change of curvature of the gripping path and gripping posture, a closed-loop mechanism from perception to judgment to action adjustment is established: when the wear of the gripper edge increases, the effective contact area and friction coefficient decrease, making it easier to produce micro-slip under the same clamping force; micro-contact pressure deviation directly reflects the deviation of the gripper force relative to the target force, and micro-slip deviation is a direct quantification of the slippage trend. The three are unified and weighted to obtain a stable contribution value, which directly indicates the reliability of the force during the gripping process; the fluctuation of the rate of change of curvature of the gripping path reflects the discrete acceleration peak and centripetal force change in the trajectory through the statistics of angular velocity and angular acceleration, thus forming a stable contribution value; the matching of the gripping posture with the reference posture quantifies the consistency of the contact geometry and the force direction and includes it in the compliance contribution value; the three are combined into an anomaly index according to preset weights, which can distinguish the anomaly types mainly due to insufficient force, dynamic fluctuation, or posture deviation. For different types, this embodiment converts the degree of anomaly into specific adjustment amounts through explicit mathematical mapping—for example, increasing or decreasing the gripper torque by the difference ratio to increase the normal force, decreasing the inertial force and trajectory amplitude by the speed reduction, and changing the orientation of the contact surface by the angle correction to improve friction conditions—and first performs simulation verification in a digital twin to avoid secondary risks, before issuing the adjustment for execution and updating the threshold and weight with real-time feedback, thereby coordinating the adjustment on three paths: mechanics, kinematics, and contact geometry, minimizing slippage, collision, and placement deviation, and improving the stability and reliability of the sorting action.

[0060] Furthermore, the mapping unit rapidly converts various composite anomaly types identified by the anomaly determination module into predefined adjustment instructions. The adjustment unit then quantitatively corrects the key operating parameters of the robotic arm according to preset adjustment coefficients and thresholds, thereby achieving coordinated optimization across the force, speed, and angle control paths: When the stability contribution value is low, the system increases the gripper torque proportionally to the stability gap to improve normal constraint and reduce micro-slippage; when the stability contribution value drops to the alarm range, the system reduces the gripping speed proportionally to the stability gap and replans the trajectory if necessary to reduce oscillations caused by inertial forces and curvature abrupt changes; when the compliance contribution value deviates from the reference, the system adjusts the placement angle or gripping posture according to the compliance difference to improve contact geometry and friction conditions; for composite anomalies, the adjustment unit performs weighted joint adjustments based on contribution value weights, with priority determined by the contribution value magnitude and preset anomaly weights. Before being issued, all proposed adjustments can be verified by parallel simulation using digital twins to check the impact of collisions and stability. The simulation results and on-site feedback are used together to correct the adjustment coefficients and thresholds according to the learning rate, so that force control, kinematic trajectory and contact geometry are matched in a closed loop in the time domain, which significantly reduces slippage, collision and placement deviation, improves the success rate of grasping and sorting efficiency, and at the same time ensures the integrity of materials and the safety of equipment. Attached Figure Description

[0061] Figure 1 This is a schematic diagram of the intelligent sorting management system for industrial robots in this embodiment;

[0062] Figure 2 This is a logic diagram for determining the stability state of the stability sensing subunit in this embodiment;

[0063] Figure 3 This is a logic diagram for determining the stability state of the stability sensing subunit in this embodiment.

[0064] Figure 4 This is a logic diagram for determining the compliance status of the compliance perception subunit in this embodiment. Detailed Implementation

[0065] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0066] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0067] Please see Figure 1 As shown, this is a schematic diagram of an intelligent sorting management system for industrial robots according to this embodiment. This embodiment provides an intelligent sorting management system for industrial robots, including:

[0068] The digital twin module is used to perform path simulation in a virtual environment based on a preset 3D robotic arm model, according to the real-time received grasping and placement task list, before physical execution, so as to generate control commands.

[0069] The data acquisition module, which is connected to the digital twin module, is used to collect in real time industrial camera images when materials placed at a preset placement angle arrive at each sorting station, micro-contact pressure deviation at the gripper of each industrial robot during the gripping process, the rate of change of curvature of the gripping path, and the micro-sliding deviation of the material.

[0070] A visual recognition module, which is connected to the data acquisition module, is used to identify the physical spatial coordinates, gripping posture, and edge wear area of ​​the material based on the image from the industrial camera.

[0071] The sorting execution module is connected to the data acquisition module, the vision recognition module and the digital twin module respectively, and is used to drive the industrial robot to perform sorting actions according to the control instructions and the physical space coordinates, so as to place the material in the target grid of the loading tray.

[0072] The sensing module is connected to the data acquisition module and the vision recognition module respectively, and is used to sense the stability state based on the micro-contact pressure deviation, the micro-slip deviation and the edge wear area, as well as to sense the motion stability state based on the rate of change of curvature of the grasping path, and to sense the compliance state based on the grasping posture.

[0073] An anomaly determination module, which is connected to the perception module, is used to determine several anomaly types based on the stability state, the motion smoothness state, and the compliance state.

[0074] An adjustment module, which is connected to the sorting execution module and the anomaly determination module respectively, is used to adjust the preset gripper torque and / or preset grasping speed and / or preset placement angle of the industrial robot according to the anomaly type and a preset adjustment mapping table.

[0075] This embodiment is applicable to intelligent operation scenarios in industrial sorting production lines, such as multi-category material sorting tasks in e-commerce warehousing, intelligent manufacturing, or flexible processing workshops. The grasping and placement task list is generated by the system based on real-time order and material information, recording in detail information such as material type, weight, size, grasping point and centroid position, current coordinates, target grid coordinates, grasping and placement posture, priority, and execution order. The virtual environment is a three-dimensional simulation space built based on digital twins, including a high-precision robotic arm model, operation scene modeling, and real-time data interface. The robotic arm model covers joint angles, link lengths, speed and torque limits, and gripper forces. The model and operation scenario modeling include workstation layout, storage compartment size, and obstacle positions. The real-time data interface can input information such as material position, posture, and motion deviation into the simulation platform. During operation, the digital twin module performs inverse kinematics solving, trajectory timing planning, and collision detection in the virtual environment based on the task list, generating optimized control commands to achieve highly synchronized sorting operations between the virtual and real worlds. This ensures that the industrial robot executes with high precision, stability, and efficiency during the grasping and placement process. The production line completes path simulation in the virtual environment in advance through the digital twin module and generates high-precision control commands to ensure that the robotic arm operates efficiently between multiple workstations.

[0076] In this embodiment, the data acquisition module achieves real-time acquisition of key parameters through a multimodal sensing system: industrial cameras are installed above and to the side of each sorting station, using high-speed, high-definition CCD or CMOS sensors to capture images of materials arriving at the station, used to identify material type, shape, and position; high-precision force sensors and micro-pressure sensing membranes are installed at the grippers to collect micro-contact pressure deviations during the gripping process in real time, with a sampling frequency of over 1kHz to capture transient force changes; the curvature change rate of the gripping path is calculated by joint angle encoders and robotic arm motion controllers, recording the changes in angular velocity and angular acceleration of each joint from the starting point to the gripping point, and calculating the curvature change rate through numerical differentiation; micro-slip deviations of the material are obtained by deploying high-resolution displacement sensors on the gripper or worktable surface or by using a visual tracking algorithm (based on continuous frame comparison of the camera to calculate the displacement of the material edge or marker point), with a resolution of up to 0.01mm; the data acquisition module transmits the above data to the sensing module after synchronizing with timestamps for subsequent stability, smoothness, and compliance analysis.

[0077] In this embodiment, the preset three-dimensional robotic arm model is a digital simulation model constructed based on the design drawings of the industrial robotic arm and the actual sensor measurement data. It adopts a kinematic framework based on Denavit-Hartenberg (DH) parameters and establishes a multi-degree-of-freedom dynamic equation system in combination with the Lagrange dynamic equations. The parameters cover the number of joints, link length, joint limit angle, maximum angular velocity and acceleration, joint friction coefficient, rated torque of the drive motor, end effector mass and grasping force characteristics. During the training process, historical running trajectory, online sensor data and high-precision simulation data are used as samples. Gradient optimization and iterative least squares method are combined to correct parameter errors, so that the model can accurately simulate the spatial position, posture changes and dynamic response of the robotic arm under multiple working conditions, thereby ensuring the accuracy and robustness of path simulation and command generation.

[0078] The preset gripper torque is the initial closing torque applied to the gripper by the industrial robot when gripping materials. Its magnitude depends on the weight, material, surface friction coefficient, and gripper structural characteristics of the material. It is typically set between 2 N·m and 15 N·m; in this embodiment, it is set to 8 N·m to ensure stable gripping while preventing material deformation or damage due to excessive torque. The preset gripping speed is the initial linear velocity of the end effector of the industrial robot during the process of gripping materials from the storage compartment to the target compartment. Its magnitude depends on the material shape, transport distance, gripping path curvature, and gripper torque. It is typically set between 0.1 m / s and 0.5 m / s; in this embodiment, it is set to 0.25 m / s to improve sorting efficiency while ensuring gripping stability and path smoothness. The preset placement angle is the tilt or rotation angle of the end effector when the industrial robot places materials into the target compartment of the pallet. Its magnitude depends on the material shape, the spatial layout of the target compartment, and material stacking requirements. The angle is usually set between 0° and 15°. In this embodiment, it is set to 5°, which can make the material be placed flat, reduce the sliding or tipping caused by improper angle, and at the same time take into account the stability and execution accuracy of the gripping action.

[0079] The digital twin module simulates the entire process of gripping and placing in advance, and the data acquisition module monitors multi-dimensional information such as gripper pressure deviation, gripping path curvature change rate, and material micro-slip deviation in real time to form a high-precision perception of the operation status. The visual recognition module ensures accurate identification of material category, grid wear status, and spatial coordinates, guaranteeing the accuracy of path planning and execution. The perception module achieves dynamic monitoring of operational risks through a comprehensive evaluation of stability, smoothness, and compliance. The anomaly determination module determines the anomaly type based on multi-parameter cross-calculation to avoid misjudgment by a single indicator. The adjustment module performs multi-parameter linkage adjustment of torque, speed, and placement angle based on the anomaly type and real-time parameter deviation, so that force control and motion control dynamically match material characteristics and operating environment, thereby significantly improving sorting efficiency and gripping stability, reducing the sorting failure rate caused by overload, posture deviation, or slippage, and achieving high-precision and robust intelligent sorting operations.

[0080] Specifically, the digital twin module includes:

[0081] The extraction unit is used to parse and extract the grasping pose, placement pose, and operation sequence of the material according to the grasping and placement task list, and obtain the parsing result;

[0082] The task compilation unit, which is connected to the extraction unit, is used to map the parsing results to the robot arm base coordinate system and the workstation coordinate system, and to generate a constrained process step sequence in combination with preset process constraints.

[0083] The simulation unit, which is connected to the task compilation unit, is used to perform path planning and motion simulation on the constrained process sequence based on a preset three-dimensional robotic arm model, so as to generate control instructions for driving the sorting execution.

[0084] Preset process constraints are operational limitations that the robotic arm must adhere to during the grasping and placement process. These constraints include the grasping sequence, working space range, obstacle avoidance distance, acceleration, and speed limits. They depend on the material size, weight, robotic arm performance, and working environment conditions. They are typically set within the range required for safe operation of the robotic arm and material integrity. In this embodiment, the grasping sequence is set to strictly follow the task list, the working space boundary is ±5mm, the acceleration is ≤1.2m / s², and the speed is ≤0.5m / s. This ensures that the robotic arm can complete the target actions smoothly, safely, efficiently, and accurately during the grasping and sorting process.

[0085] By extracting the material gripping pose, placement pose, and operation sequence from the gripping and placement task list, and mapping the analysis results to the robot arm base coordinate system and the workstation coordinate system, and combining them with preset process constraints to generate a constrained sequence of steps, path planning and motion simulation can be performed in a virtual environment based on a preset 3D robot arm model, thereby generating precise control commands. This process achieves dynamic matching and optimization between gripping torque, gripping speed, and placement angle by coordinating the material spatial position, gripping posture, path curvature, and robot arm motion parameters. This enables the robot arm to grip stably, move smoothly, and place accurately during physical execution, improving overall sorting efficiency and operational reliability.

[0086] Specifically, the simulation unit includes:

[0087] The parameter determination subunit is used to perform inverse kinematics solution and time parameterized trajectory planning on the constrained work step sequence based on the preset three-dimensional robotic arm model in a virtual environment, so as to obtain the expected position, velocity and acceleration of each joint and determine the motion parameters.

[0088] The trajectory timing generation subunit is connected to the parameter determination subunit to generate a complete joint space running trajectory and end effector action timing based on the motion parameters.

[0089] A constraint subunit, which is connected to the trajectory timing generation subunit, is used to perform full-process obstacle avoidance and collision detection according to the preset work station layout and preset obstacle modeling information, so as to constrain the joint space running trajectory and obtain the constrained running trajectory.

[0090] The instruction generation subunit is connected to the trajectory timing generation subunit and the constraint subunit respectively, and is used to package the constrained running trajectory and the end effector action timing according to the controller communication protocol to generate the control instruction.

[0091] In this embodiment, the preset workstation layout refers to the fixed coordinates and arrangement of each work area, pallet, conveyor belt, workbench, and material placement position in three-dimensional space when the industrial robot performs sorting tasks. These coordinates and spatial relationships constitute a reference frame for the robotic arm when performing grasping and placing actions, guiding path planning and enabling the robotic arm to move efficiently between workstations without deviating from the predetermined work area. The preset obstacle modeling information refers to the three-dimensional modeling of fixed or movable obstacles (such as supports, sensors, grippers, other robots, or unplaced materials) that may exist in the work environment, and recording their spatial position, size, and shape information in the virtual environment. This information is used for collision detection and obstacle avoidance during trajectory planning, ensuring that the robotic arm does not come into contact with obstacles when performing actions, thereby guaranteeing operational safety and accuracy.

[0092] By performing inverse kinematics solving and time-parameterized trajectory planning on the constrained work step sequence, the expected position, velocity, and acceleration of each joint at different time points are accurately calculated, thus forming a complete joint spatial running trajectory and end effector action sequence. Combined with the preset workstation layout and obstacle modeling information, the system can constrain the trajectory based on obstacle avoidance and collision detection throughout the process, ensuring that the robotic arm maintains stable movement in actual operation and can efficiently and accurately complete the grasping and placement tasks. At the same time, by packaging the trajectory and action sequence according to the controller protocol to generate control instructions, a high consistency between the virtual environment and physical execution is achieved, ensuring the safety and reliability of the sorting action.

[0093] Specifically, the sensing module includes:

[0094] A stable sensing unit is used to calculate a stability index based on the micro-contact pressure deviation, the micro-slip deviation, and the edge wear area within a preset sensing time, and to sense the stability state based on the stability index.

[0095] A stability sensing unit is used to calculate a stability index based on the rate of change of curvature of the grasping path within the same preset sensing time, and to sense the stability state based on the stability index.

[0096] The compliance perception unit is used to calculate a compliance index based on the posture and a preset posture library within the same preset perception time period, and to perceive the compliance status based on the compliance index.

[0097] The preset sensing duration is the time window during which the system collects data on micro-contact pressure deviation, micro-slippage deviation, rate of change of curvature of the grasping path, and grasping posture during the grasping action. It depends on the grasping cycle of the robotic arm, the size and weight of the material, and the sampling frequency of the industrial camera and sensors. It is usually set between 0.5 seconds and 2 seconds. In this embodiment, it is set to 1 second, which can realize real-time dynamic monitoring of the grasping process while ensuring data sufficiency, thereby providing a reliable data foundation for the accurate calculation of stability, smoothness, and compliance indices.

[0098] By comprehensively analyzing micro-contact pressure deviation, micro-slippage deviation, and wear area at the edge of the gripper, the stability of materials during the gripping process is quantified. The smoothness of the robotic arm's movement is assessed through the rate of change of curvature along the gripping path. Simultaneously, a compliance index is calculated by comparing the gripping posture with a preset posture library, thus forming a multi-dimensional state perception of the gripping process. This multi-parameter joint analysis enables the system to accurately reflect the real-time performance of the robotic arm in terms of material force, motion trajectory, and gripping posture during actual operation. Through the mutual mapping and weighting of stability, smoothness, and compliance indices, it can effectively guide the subsequent optimization and adjustment of gripper torque, gripping speed, and placement angle, achieving dynamic monitoring and intelligent control of the sorting process, improving gripping success rate, and reducing the risk of material slippage and damage.

[0099] Please see Figure 2 As shown, this is a logic diagram for determining the stability state of the stability sensing subunit in this embodiment. In this embodiment, the stability sensing subunit includes:

[0100] The stabilization subunit is used to perform maximum-minimum normalization processing on the micro-contact pressure deviation at the end of the preset sensing time based on all the micro-contact pressure deviations within the preset sensing time to obtain a pressure deviation normalization value; and to perform maximum-minimum normalization processing on the micro-slip deviation at the end of the preset sensing time based on all the micro-slip deviations within the preset sensing time to obtain a slip deviation normalization value; and to perform maximum-minimum normalization processing on the edge wear area at the end of the preset sensing time based on all the edge wear areas within the preset sensing time to obtain a wear area normalization value.

[0101] The stability index calculation subunit, which is connected to the stability normalization subunit, is used to perform a weighted summation calculation on the pressure deviation normalization value, the preset pressure deviation weight, the sliding deviation normalization value, the preset sliding deviation weight, the wear area normalization value, and the preset wear area weight to obtain the stability index.

[0102] A stability sensing subunit, which is connected to the stability index calculation subunit, is used to determine that the stability state is sensed when the stability index is greater than a preset stability index threshold.

[0103] The preset pressure deviation weight is used to represent the degree of influence of micro-contact pressure deviation on the stability index. It depends on the material properties, gripper type, and gripping task requirements, and is usually set between 0 and 1. In this embodiment, it is set to 0.4, which can reasonably reflect the contribution of pressure deviation to gripping stability when calculating the stability index, thereby optimizing gripper torque adjustment. The preset slip deviation weight is used to represent the degree of influence of micro-slippage of the material during the gripping process on the stability index. It depends on the material surface friction coefficient, gripping speed, and gripper surface material, and is usually set between 0 and 1. In this embodiment, it is set to 0.35, which can accurately reflect the influence of slip deviation on the material gripping stability in the stability index calculation, assisting in gripping speed optimization. The preset wear area weight is used to represent the weight of the material... The impact of grid edge wear on the stability index depends on the frequency of grid use, material stacking method, and the precision of the gripping machinery. It is usually set between 0 and 1. In this embodiment, it is set to 0.25, which can reasonably assess the comprehensive impact of wear on gripping stability in the stability index calculation, ensuring the reliability of the robotic arm's actions during sorting. The preset stability index threshold is used to determine whether the industrial robot maintains sufficient stability during gripping. It depends on the material type, gripper characteristics, and safety requirements of the gripping task. It is usually set between 0 and 1. In this embodiment, it is set to 0.7, which can effectively identify the stable state of the gripping action when the stability index exceeds this threshold, thereby guiding the reasonable adjustment of gripper torque and gripping strategy, ensuring a smooth and reliable material sorting process.

[0104] By normalizing the micro-contact pressure deviation, micro-slip deviation, and wear area at the edge of the storage compartment within a preset sensing time, various raw data are mapped to a unified dimensional range to eliminate the influence of differences in the scale of different parameters. Subsequently, a stability index is calculated by weighted summation based on the preset weights of each parameter. This index comprehensively reflects the mechanical response, material sliding trend, and wear of the contact surface during the robotic arm's grasping process. Thus, when the stability index exceeds a preset threshold, the stability state can be accurately determined, enabling dynamic perception of the mechanical properties of the grasping action and the state of the material. This ensures that the control accuracy and safety of the industrial robot in the sorting process are effectively improved.

[0105] Please see Figure 3 As shown, this is a logic diagram for determining the stability state of the stability sensing subunit in this embodiment. In this embodiment, the stability sensing subunit includes:

[0106] The steady-state normalization subunit is used to calculate the standard deviation of the rate of change of curvature of all the grasping paths from the initial time to each time within the same preset perception time, to obtain several curvature change fluctuation values, and to perform maximum-minimum normalization processing on each curvature change fluctuation value to obtain several fluctuation normalization values.

[0107] A stability index calculation subunit, which is connected to the stability normalization subunit, is used to calculate the average value of all the fluctuation normalization values ​​to obtain the stability index.

[0108] A stability sensing subunit, which is connected to the stability index calculation subunit, is used to determine that the stability state is sensed when the stability index is greater than a preset stability index threshold.

[0109] The preset stability index threshold is a critical value used to determine whether the robotic arm's grasping action maintains stable movement. Its value depends on the required grasping accuracy, material characteristics, and the dynamic response capability of the robotic arm itself. It is usually set between 0 and 1. In this embodiment, it is set to 0.75. When the stability index is higher than this threshold, the robotic arm's movement can be determined to be stable, thereby ensuring that vibration during material grasping and placement is minimized, reducing the risk of slippage or deviation, and improving the reliability and efficiency of sorting actions.

[0110] By normalizing the standard deviation of the rate of change of curvature of the grasping path within a preset sensing time, the curvature change fluctuation value is calculated, and the average value is further calculated to obtain the stability index. This stability index reflects the smoothness of the robotic arm's movement during the grasping process. By comparing the stability index with a preset stability index threshold, the system can accurately determine whether the movement is stable, thereby guiding the optimization of grasping speed and trajectory. This reduces vibration and deviation when the robotic arm performs sorting actions, ensuring smooth sliding of materials during grasping and placement, and improving overall sorting efficiency and reliability.

[0111] Please see Figure 4 As shown, this is a logic diagram for determining the compliance status of the compliance perception subunit in this embodiment. In this embodiment, the compliance perception subunit includes:

[0112] The matching degree calculation subunit is used to compare the grasping posture at each moment within the same preset perception time with the reference posture in the preset posture library to obtain several matching degrees.

[0113] A compliance index calculation subunit, which is connected to the matching degree calculation subunit, is used to calculate the average value of all the matching degrees to obtain the compliance index;

[0114] A compliance perception subunit, which is connected to the compliance index calculation subunit, is used to determine the perceived compliance status when the compliance index is greater than a preset compliance index threshold.

[0115] In this embodiment, within a preset sensing time period, the system records the spatial pose of the industrial robot gripper at each sampling moment in real time at a fixed sampling frequency, including position coordinates and attitude angles. Simultaneously, a preset attitude library stores a sequence of reference attitudes corresponding to the grasping action of the target material. The matching degree calculation subunit compares the grasping attitude collected at each moment with the corresponding reference attitude in the preset attitude library one by one. By calculating indicators such as displacement difference and rotation angle difference, it obtains the matching degree value at each moment. Then, the matching degree values ​​at all moments are used for subsequent compliance index calculation, thereby comprehensively reflecting the accuracy and stability of the grasping action throughout the entire sensing cycle.

[0116] The preset compliance index threshold is a standard value used to determine whether the grasping posture is compliant. It depends on the accuracy requirements of the robot's grasping action, the shape of the material, and the complexity of the grasping environment. It is usually set between 0.7 and 0.95. In this embodiment, it is set to 0.85, which can effectively distinguish abnormal grasping behavior while ensuring the stability of the grasping action, and provide an accurate basis for subsequent adjustment of the gripper torque, grasping speed, or placement angle.

[0117] By comparing the gripping posture at each moment during the gripping process with reference postures in a preset posture library, the matching degree is obtained, and the average of all matching degrees is calculated to form a compliance index, which reflects the accuracy of the gripping action in terms of spatial position and posture. When the compliance index is higher than a preset threshold, the system determines that the gripping posture is compliant, thereby ensuring the accuracy of the relative position and angle between the gripper and the material, reducing tilting or rotation errors. At the same time, in conjunction with stability and smoothness parameters, it helps to achieve safe gripping, smooth transportation and accurate placement of materials, improving overall sorting efficiency and reliability.

[0118] Specifically, the anomaly determination module includes:

[0119] The contribution value calculation unit is used to calculate a stable contribution value based on a stability index, a stable contribution value based on a stability index, and a compliance contribution value based on a compliance index when the stability state, the stability state, and the compliance state are perceived.

[0120] Where Q1 = (1-q1)×a1, Q2 = (1-q2)×a2, Q3 = (1-q3)×a3, Q1 is the stable contribution value, Q2 is the stable contribution value, Q3 is the compliant contribution value, a1 is the preset stable contribution weight, a2 is the preset stable contribution weight, a3 is the preset compliant contribution weight, q1 is the stability index, q2 is the stable index, and q3 is the compliance index;

[0121] An anomaly index calculation unit is used to calculate an anomaly index based on the stable contribution value, the steady contribution value, and the compliant contribution value.

[0122] Where Y = Q1×i1 + Q2×i2 + Q3×i3, Y is the anomaly index, i1 is the preset anomaly stability weight, i2 is the preset anomaly stability weight, and i3 is the preset anomaly compliance weight.

[0123] The type determination unit is used to determine several of the anomaly types based on the anomaly index, the stable contribution value, the steady contribution value, and the compliance contribution value.

[0124] By converting the stability index, smoothness index, and compliance index into corresponding contribution values, and then comprehensively calculating the anomaly index based on preset weights, the potential anomalies in the sorting process can be quantitatively determined. The stability contribution value, smoothness contribution value, and compliance contribution value reflect the force control accuracy, motion smoothness, and compliance degree of material gripping, respectively. The anomaly index, on the other hand, integrates the weighted influence of each contribution value, which can clearly distinguish different types of abnormal states and provide a scientific basis for subsequent adjustments to gripper torque, gripping speed, and placement angle. This enables dynamic optimization and reliability improvement of robot actions in complex sorting environments.

[0125] The preset stability contribution weight is a weight used to reflect the degree of influence of stability on the overall anomaly index in anomaly determination. It depends on the gripper torque, micro-contact pressure deviation and material sliding characteristics during material gripping. It is usually set between 0 and 1. In this embodiment, it is set to 0.4, which can ensure that stability changes have an appropriate contribution to the anomaly index, so that anomaly determination can sensitively reflect gripping force control problems.

[0126] The preset stability contribution weight is a weight used to reflect the degree of influence of motion stability on the overall anomaly index in anomaly judgment. It depends on the rate of change of curvature of the grasping path and the fluctuation of motion speed. It is usually set between 0 and 1. In this embodiment, it is set to 0.35, which enables the anomaly index to accurately reflect the deviation of the robot arm's motion stability, thereby capturing situations that may cause material swaying.

[0127] The preset compliance contribution weight is a weight used to reflect the degree of influence of the compliance of the grasping posture on the overall anomaly index in anomaly judgment. It depends on the matching degree between the grasping posture and the preset posture library, and is usually set between 0 and 1. In this embodiment, it is set to 0.25, which can ensure that the posture deviation has a reasonable impact on the anomaly index and keep the grasping action in line with the design specifications. The preset anomaly stability weight is a parameter used to reflect the weight of the stability contribution value on the final anomaly index when calculating the anomaly index. It depends on the system's sensitivity to stability anomalies and is usually set between 0 and 1. In this embodiment, it is set to 0.5, which can reasonably amplify or adjust the impact of stability anomalies in the anomaly index. The preset anomaly stability weight is a parameter used to reflect the weight of the stability contribution value to the final anomaly index when calculating the anomaly index. It depends on the system's sensitivity to motion stability anomalies and is usually set between 0 and 1. In this embodiment, it is set to 0.3, which can make the impact of motion fluctuations on anomaly judgment moderate. The preset anomaly compliance weight is a parameter used to reflect the weight of the compliance contribution value to the final anomaly index when calculating the anomaly index. It depends on the system's sensitivity to grasping posture deviation anomalies and is usually set between 0 and 1. In this embodiment, it is set to 0.2, which can make posture anomalies make a reasonable contribution to the final anomaly index without masking stability and stability anomalies.

[0128] By converting the stability index, smoothness index, and compliance index into corresponding contribution values, and then comprehensively calculating the anomaly index based on preset weights, the system achieves quantitative judgment of potential anomalies during the sorting process. The stability contribution value, smoothness contribution value, and compliance contribution value reflect the precision of force control, motion smoothness, and compliance degree of grasping posture, respectively. The anomaly index, by comprehensively considering the weighted influence of each contribution value, can clearly distinguish different types of abnormal states and provide a scientific basis for subsequent adjustments to gripper torque, grasping speed, and placement angle. This enables dynamic optimization and reliability improvement of robot actions in complex sorting environments.

[0129] Specifically, the type determination unit is used to determine the anomaly type as the first type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is less than a preset stable contribution threshold, the steady contribution value is greater than a preset steady contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold.

[0130] The type determination unit is further configured to determine the anomaly type as the second type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is greater than or equal to a preset stable contribution threshold, the steady contribution value is less than or equal to a preset steady contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold.

[0131] The type determination unit is further configured to determine that the anomaly type is a third type when the anomaly index is greater than a preset anomaly index threshold, the stable contribution value is less than a preset stable contribution threshold, the steady contribution value is less than or equal to a preset steady contribution threshold, and the compliance contribution value is less than or equal to a preset compliance contribution threshold.

[0132] The preset anomaly index threshold is the critical value at which the system judges whether the overall anomaly intensity should trigger intervention. It depends on the magnitude of the contribution values ​​of the three paths: stable, steady, and compliant, the weight of each contribution, and the system's allowable error tolerance. It is generally recommended to set it in the range of 0.05 to 0.20 to balance false alarms and false negatives. In this embodiment, it is set to 0.12, which can promptly identify significant deviations caused by combined anomalies of force, motion, or attitude above common operational fluctuations, thereby triggering subsequent parameter adjustments or simulation verification, and avoiding small fluctuations being misjudged as serious faults or major deviations being missed. A preset stability contribution threshold is used to determine when the stability contribution value indicates insufficient gripping force reliability. It depends on the statistical distribution of the stability contribution weight and stability index (determined by the material, fixture, and gripping process). It is generally recommended to set it between 0.05 and 0.25 to adapt to different materials and gripping strategies. In this embodiment, it is set to 0.12, which can separate gripping instability caused by reduced gripper contact area, decreased friction, or contact force deviation from general fluctuations, thereby preferentially triggering incremental adjustment of gripper torque to restore normal force and prevent micro-slippage. A preset smoothness contribution threshold is used to determine when motion smoothness reaches a level requiring intervention. It depends on the smoothness contribution weight, the time-domain characteristics of the rate of curvature change, and the system's tolerance to trajectory amplitude. It is generally recommended to set it in the range of 0.04 to 0.18. In this embodiment... Setting it to 0.09 allows for the differentiation between occasional jitter and systematic instability requiring speed / trajectory adjustment when short-term large fluctuations or frequent fluctuations in trajectory curvature are detected. This guides the reduction of gripping speed or trajectory replanning to reduce swaying caused by inertial forces. The preset compliance contribution threshold is used to determine whether the posture matching deviation has reached the level requiring correction. It depends on the compliance contribution weight, the strictness of the preset posture library, and the sensitivity of the material to the placement angle. It is generally recommended to set it between 0.02 and 0.12 to accommodate different accuracy requirements. In this embodiment, it is set to 0.04, which can trigger quantitative correction of the placement angle or gripping posture when a systematic deviation between the gripping posture and the reference posture is detected (e.g., significant inconsistency in the contact surface direction). This improves the contact geometry and friction conditions and reduces the risk of overturning or slippage.

[0133] By quantifying and integrating mechanical and surface parameters such as micro-contact pressure deviation, material micro-slippage deviation, and wear area of ​​the gripper edge with kinematic parameters such as the rate of change of curvature of the gripping path and gripping posture, a closed-loop mechanism from perception to judgment to action adjustment is established: When the wear of the gripper edge increases, the effective contact area and friction coefficient decrease, making it easier to produce micro-slippage under the same clamping force; micro-contact pressure deviation directly reflects the deviation of the gripper force relative to the target force, and micro-slippage deviation is a direct quantification of the slippage trend. The three are unified and weighted to obtain a stable contribution value, which directly indicates the reliability of the force during the gripping process; the fluctuation of the rate of change of curvature of the gripping path reflects the discrete acceleration peak and centripetal force change in the trajectory through the statistics of angular velocity and angular acceleration, thus forming a stable contribution value; the matching of the gripping posture with the reference posture quantifies the consistency of the contact geometry and the force direction and includes it in the compliance contribution value; the three are combined into an anomaly index according to preset weights, which can distinguish the anomaly types mainly due to insufficient force, dynamic fluctuation, or posture deviation. For different types, this embodiment converts the degree of anomaly into specific adjustment amounts through explicit mathematical mapping—for example, increasing or decreasing the gripper torque by the difference ratio to increase the normal force, decreasing the inertial force and trajectory amplitude by the speed reduction, and changing the orientation of the contact surface by the angle correction to improve friction conditions—and first performs simulation verification in a digital twin to avoid secondary risks, before issuing the adjustment for execution and updating the threshold and weight with real-time feedback, thereby coordinating the adjustment on three paths: mechanics, kinematics, and contact geometry, minimizing slippage, collision, and placement deviation, and improving the stability and reliability of the sorting action.

[0134] Specifically, the adjustment module includes:

[0135] The mapping unit is used to map the first type, the second type and the third type to the preset adjustment mapping table respectively, so as to obtain the first adjustment instruction corresponding to the first type, the second adjustment instruction corresponding to the second type and the third adjustment instruction corresponding to the third type;

[0136] An adjustment unit, connected to the mapping unit, is used to adjust the preset gripper torque according to a preset torque adjustment coefficient based on the first adjustment command, and to adjust the preset gripping speed according to a preset speed adjustment coefficient based on the second adjustment command, wherein F'=F×k1, V'=V×k2, F' is the adjusted preset gripper torque, F is the original preset gripper torque, k1 is the preset torque adjustment coefficient, V' is the adjusted preset gripping speed, V is the original preset gripping speed, and k2 is the preset speed adjustment coefficient;

[0137] The adjustment unit is further configured to, based on the third adjustment instruction, adjust the preset gripper torque according to the preset stable contribution threshold and the stable contribution value, adjust the preset gripping speed according to the preset smooth contribution threshold and the stable contribution value, and adjust the preset placement angle according to the compliance contribution value and the preset compliance contribution threshold, wherein F'=F×[1+k3×(Q1'-Q1) / Q1], V'=V×[1-k4×(Q2-Q2') / Q2'], R'=R×[1-k5×(Q3'-Q3) / Q3], k3 is the preset torque correction coefficient, k4 is the preset speed correction coefficient, R' is the preset placement angle before adjustment, R is the preset placement angle after adjustment, k5 is the preset angle correction coefficient, Q1' is the preset stable contribution threshold, Q2' is the preset smooth contribution threshold, and Q3' is the preset compliance contribution threshold.

[0138] The preset torque adjustment factor is a multiplier factor used to proportionally amplify or reduce the preset gripper torque. It depends on the robotic arm's driving capability, gripper mechanism rigidity, material fragility, and the maximum allowable gripping force. It is typically set between 1.01 and 1.20 (greater than 1 indicates amplification). In this embodiment, it is set to 1.05, which allows for a small-step increase in gripping torque when slight instability is detected, quickly improving normal constraint to suppress micro-slippage, while avoiding material deformation or exceeding the actuator's safe range due to a large one-time force increase. The preset speed adjustment factor is a multiplier factor used to proportionally scale the preset gripping speed, depending on the robotic arm's dynamics. The response, working distance, trajectory curvature, and material sensitivity to acceleration are typically set between 0.70 and 0.95 (less than 1 indicates deceleration). In this embodiment, it is set to 0.90, which can moderately reduce the end velocity when unstable motion is detected, reducing oscillations caused by inertial forces and sudden changes in trajectory curvature, thereby improving motion smoothness and reducing the need for secondary adjustments. The preset torque correction coefficient is a learning rate / amplification coefficient used to proportionally correct the gripper torque according to the difference in stable contribution. It depends on the tolerance of field testing, the feedback accuracy of the gripping mechanism, and the noise level of the force sensor; it is typically set between 0.1 and 0.8 (the larger the value, the more aggressive the response, but...). (The higher the risk), the higher the value of the torque compensation. In this embodiment, it is set to 0.50, which can provide obvious but controlled torque compensation when the stability is insufficient, so that the force is restored to the safe range. At the same time, it is used in conjunction with the stepping / simulation verification strategy to avoid overshoot or the introduction of new problems. The preset speed correction coefficient is a learning rate / attenuation coefficient used to adjust the grasping speed proportionally according to the difference in stability contribution. It depends on the braking and re-acceleration capabilities of the robotic arm, the process delay, and the throughput requirements. It is usually set between 0.1 and 0.8. In this embodiment, it is set to 0.60, which can decelerate at an appropriate ratio when an abnormality in stability is detected, quickly suppressing the trajectory amplitude and acceleration peak, thereby minimizing the impact on stability. While reducing efficiency loss, it significantly reduces the risk of swaying and falling off caused by speed. The preset angle correction coefficient is a proportional factor used to map the difference in compliance contribution to the amount of placement angle correction. It depends on the sensitivity of the material to the placement angle, the space margin of the target grid, and the accuracy of the fixture's diagonal adjustment. It is usually set between 0.05 and 1.0 (the larger the value, the stronger the response to angle deviation). In this embodiment, it is set to 0.80, which can quickly adjust the placement angle to improve the contact geometry and friction conditions when compliance deviation is detected, thereby significantly reducing the probability of tilting, misalignment, or slippage. At the same time, it is combined with the maximum angle limit to avoid affecting subsequent stacking.

[0139] The mapping unit rapidly converts various composite anomaly types identified by the anomaly determination module into predefined adjustment instructions. The adjustment unit then quantitatively corrects the key operating parameters of the robotic arm according to preset adjustment coefficients and thresholds, thereby achieving coordinated optimization across the force, speed, and angle control paths: When the stability contribution value is low, the system increases the gripper torque proportionally to the stability gap to improve normal constraint and reduce micro-slippage; when the stability contribution value drops to the alarm range, the system reduces the gripping speed proportionally to the stability gap and replans the trajectory if necessary to reduce oscillations caused by inertial forces and curvature abrupt changes; when the compliance contribution value deviates from the reference, the system adjusts the placement angle or gripping posture according to the compliance difference to improve contact geometry and friction conditions; for composite anomalies, the adjustment unit performs weighted joint adjustments based on contribution value weights, with priority determined by the contribution value magnitude and preset anomaly weights. Before being issued, all proposed adjustments can be verified by parallel simulation using digital twins to check the impact of collisions and stability. The simulation results and on-site feedback are used together to correct the adjustment coefficients and thresholds according to the learning rate, so that force control, kinematic trajectory and contact geometry are matched in a closed loop in the time domain, which significantly reduces slippage, collision and placement deviation, improves the success rate of grasping and sorting efficiency, and at the same time ensures the integrity of materials and the safety of equipment.

[0140] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent sorting management system for an industrial robot, characterized in that, The method comprises the following steps: a digital twin module is used to simulate a path in a virtual environment based on a preset three-dimensional robot model according to a real-time received pick-and-place task list before physical execution, to generate control instructions; a data acquisition module is used to acquire industrial camera images when a material placed at a preset placement angle reaches each sorting station, micro-contact pressure deviation of gripper during the pick-up process, pick-up path curvature change rate, and micro-sliding deviation of the material; a visual recognition module is used to recognize the physical space coordinates, pick-up posture, and edge wear area of the material according to the industrial camera images; a sorting execution module is used to drive the industrial robot to perform sorting actions according to the control instructions and the physical space coordinates, to place the material in the target grid of the carrier plate; a perception module is used to perceive stability state according to the micro-contact pressure deviation, the micro-sliding deviation, and the edge wear area, to perceive motion stability state according to the pick-up path curvature change rate, and to perceive compliance state according to the pick-up posture; an abnormality determination module is used to determine several abnormality types according to the stability state, the motion stability state, and the compliance state; an adjustment module is used to adjust the preset gripper torque and / or the preset pick-up speed and / or the preset placement angle of the industrial robot according to the abnormality types and a preset adjustment mapping table; The perception module comprises: a stability perception unit is used to calculate a stability index according to the micro-contact pressure deviation, the micro-sliding deviation, and the edge wear area within a preset perception time length, and to perceive the stability state according to the stability index; a smooth perception unit is used to calculate a smooth index according to the pick-up path curvature change rate within the same preset perception time length, and to perceive the motion stability state according to the smooth index; a compliance perception unit is used to calculate a compliance index according to the posture and a preset posture library within the same preset perception time length, and to perceive the compliance state according to the compliance index; The stability perception unit comprises: a stability normalization sub-unit is used to perform maximum-minimum value normalization processing on the micro-contact pressure deviation at the end of the preset perception time length according to all the micro-contact pressure deviations within the preset perception time length, to obtain a pressure deviation normalized value, to perform maximum-minimum value normalization processing on the micro-sliding deviation at the end of the preset perception time length according to all the micro-sliding deviations within the preset perception time length, to obtain a sliding deviation normalized value, and to perform maximum-minimum value normalization processing on the edge wear area at the end of the preset perception time length according to all the edge wear areas within the preset perception time length, to obtain a wear area normalized value. A stability index calculation subunit connected with the stability normalization subunit, configured to calculate a weighted sum of the pressure deviation normalized value, a preset pressure deviation weight, the sliding deviation normalized value, a preset sliding deviation weight, the wear area normalized value, and a preset wear area weight, to obtain the stability index, wherein the preset pressure deviation weight is 0.4, the preset sliding deviation weight is 0.35, and the preset wear area weight is 0.25; A stability perception subunit connected with the stability index calculation subunit, configured to determine that the stability state is perceived when the stability index is greater than a preset stability index threshold value; The smoothness perception unit comprises: A smoothness normalization subunit, configured to calculate a standard deviation of all the curvature change rates of the grabbing path from an initial time to each time within the preset perception time length, to obtain a plurality of curvature change fluctuation values, and to perform maximum-minimum value normalization processing on each curvature change fluctuation value, to obtain a plurality of fluctuation normalized values; A smoothness index calculation subunit connected with the smoothness normalization subunit, configured to calculate an average value of all the fluctuation normalized values, to obtain the smoothness index; A smoothness perception subunit connected with the smoothness index calculation subunit, configured to determine that the smoothness state is perceived when the smoothness index is greater than a preset smoothness index threshold value; The compliance perception unit comprises: A matching degree calculation subunit, configured to compare the grabbing posture at each time within the preset perception time length with a reference posture in a preset posture library, to obtain a plurality of matching degrees; A compliance index calculation subunit connected with the matching degree calculation subunit, configured to calculate an average value of all the matching degrees, to obtain the compliance index; A compliance perception subunit connected with the compliance index calculation subunit, configured to determine that the compliance state is perceived when the compliance index is greater than a preset compliance index threshold value; The anomaly determination module comprises: A contribution value calculation unit, configured to calculate a stability contribution value according to the stability index when the stability state, the smoothness state, and the compliance state are perceived, to calculate a smoothness contribution value according to the smoothness index, and to calculate a compliance contribution value according to the compliance index; An anomaly index calculation unit, configured to calculate an anomaly index according to the stability contribution value, the smoothness contribution value, and the compliance contribution value; A type determination unit, configured to determine a plurality of the anomaly types according to the anomaly index, the stability contribution value, the smoothness contribution value, and the compliance contribution value.

2. The intelligent sorting management system for an industrial robot according to claim 1, characterized by, The digital twin module comprises: An extraction unit, configured to parse and extract a grabbing pose, a placing pose, and an operation sequence of the material according to the grabbing and placing task list, to obtain a parsing result; A task compilation unit connected with the extraction unit, configured to map the parsing result to a robot base coordinate system and a station coordinate system, and to generate a constraint work step sequence in combination with a preset process constraint; A simulation unit connected with the task compilation unit, configured to perform path planning and motion simulation on the constraint work step sequence based on a preset three-dimensional robot model, to generate a control instruction for driving sorting execution.

3. The intelligent sorting management system for industrial robots according to claim 2, characterized in that, The simulation unit comprises: The parameter determination subunit is configured to determine motion parameters by performing inverse kinematics solving and time parameterized trajectory planning on the constraint process sequence based on the preset three-dimensional robot model in the virtual environment, to obtain expected positions, velocities, and accelerations of each joint. The trajectory timing generation subunit is connected with the parameter determination subunit and configured to generate complete joint space running trajectories and end effector action timing according to the motion parameters. The constraint subunit is connected with the trajectory timing generation subunit and configured to perform full-range obstacle avoidance and collision detection according to preset work station layout and preset obstacle modeling information, to constrain the joint space running trajectories, to obtain constrained running trajectories. The instruction generation subunit is connected with the trajectory timing generation subunit and the constraint subunit respectively, and configured to package the constrained running trajectories and the end effector action timing according to a controller communication protocol to generate the control instructions.

4. The intelligent sorting management system for industrial robots according to claim 3, characterized in that, The type determination unit is configured to determine that the abnormal type is a first type when the abnormality index is greater than a preset abnormality index threshold, the stability contribution value is less than a preset stability contribution threshold, the smoothness contribution value is greater than a preset smoothness contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold. The type determination unit is further configured to determine that the abnormal type is a second type when the abnormality index is greater than a preset abnormality index threshold, the stability contribution value is greater than or equal to a preset stability contribution threshold, the smoothness contribution value is less than or equal to a preset smoothness contribution threshold, and the compliance contribution value is greater than a preset compliance contribution threshold. The type determination unit is further configured to determine that the abnormal type is a third type when the abnormality index is greater than a preset abnormality index threshold, the stability contribution value is less than a preset stability contribution threshold, the smoothness contribution value is less than or equal to a preset smoothness contribution threshold, and the compliance contribution value is less than or equal to a preset compliance contribution threshold.

5. The intelligent sorting management system for industrial robots according to claim 4, characterized in that, The adjustment module comprises: The mapping unit is configured to map the first type, the second type, and the third type to the preset adjustment mapping table respectively, to obtain a first adjustment instruction corresponding to the first type, a second adjustment instruction corresponding to the second type, and a third adjustment instruction corresponding to the third type. The adjustment unit is connected with the mapping unit and configured to adjust the preset jaw torque according to a preset torque adjustment coefficient based on the first adjustment instruction, and adjust the preset grabbing speed according to a preset speed adjustment coefficient based on the second adjustment instruction. The adjustment unit is further configured to adjust the preset jaw torque according to the preset stability contribution threshold and the stability contribution value, adjust the preset grabbing speed according to the preset smoothness contribution threshold and the stability contribution value, and adjust the preset placement angle according to the compliance contribution value and the preset compliance contribution threshold based on the third adjustment instruction.

Images