A Collaborative Control System and Method for Belt Sanders for Grinding and Polishing Robots

Abstract

This invention relates to the technical field of automated grinding and polishing with industrial robots, and discloses a collaborative control system and method for belt sanders used in grinding and polishing robots. The collaborative control method is applied to communication and collaborative equipment and specifically includes the following steps: S11: Establishing a real-time communication network based on EtherCAT and Modbus, connecting the robot, belt sander, PLC controller, multi-dimensional force sensor, and 3D vision module to achieve data synchronization and command alignment, generating a unified underlying data platform; S12: Based on the real-time communication network of the underlying data platform, identifying workpiece surface features through the 3D vision module and calling the process database. This invention achieves adaptive matching of belt sander speed and robot trajectory to the workpiece surface through a force-position coupling dynamic collaborative mechanism, overcoming the problems of over-grinding, under-grinding, and uneven surface quality caused by pressure and speed mismatch in grinding complex curved surfaces.

Description

Technical Field

[0001] This invention relates to the technical field of automated grinding and polishing for industrial robots, and more specifically, to a collaborative control system and method for belt sanders for grinding and polishing robots. Background Technology

[0002] Currently, in high-end precision manufacturing fields such as aerospace, automobile manufacturing, and energy equipment, extremely high requirements are placed on the surface polishing quality and processing consistency of complex curved metal components. The use of industrial robots equipped with polishing tools for automated operations has become an important development trend. Among them, belt polishing is widely used due to its elastic contact, good heat dissipation, and adaptability to complex surfaces.

[0003] Existing technical solutions typically integrate industrial robots and belt sanders as two relatively independent units. The robot is responsible for holding the workpiece or tool to complete the trajectory movement, while the belt sander is controlled by an independent frequency converter to start, stop and speed of its motor. The two are linked through simple digital I / O signals or basic communication protocols. At the control layer, most systems use position control based on preset trajectories or introduce single-dimensional multi-dimensional force sensors to achieve constant force grinding. However, the motion planning of the robot and the operating parameters of the belt sander are often preset and relatively fixed, lacking the ability to dynamically and collaboratively adjust according to the real-time shape of the workpiece and the grinding status.

[0004] The following are specific technical problems that existing technologies face when dealing with the demands of high precision, high efficiency, and high flexibility in production: 1. Because the operating parameters of the belt sander are not dynamically matched with the robot's motion trajectory and the surface features of the workpiece, when grinding complex curved surfaces, areas with sudden changes in curvature are prone to over-grinding, under-grinding, or vibration marks due to mismatch in contact pressure or linear velocity. 2. The single-station operation mode means that the belt sander is in a waiting state when the robot is loading or unloading or when the workpiece is being switched. The overall utilization rate of the equipment is low. Moreover, for new workpieces, a lot of time is required for robot trajectory teaching, belt sander parameter exploration and force control parameter tuning. The debugging cycle is long and the cost is high.

[0005] Therefore, there is an urgent need for intelligent control methods and systems that can achieve deep collaboration between grinding and polishing robots and belt sanders in terms of trajectory, speed, force control, and process. Summary of the Invention

[0006] The purpose of this invention is to provide a collaborative control system and method for belt sanders for grinding and polishing robots. Through a dynamic collaborative mechanism of force-position coupling, the speed of the belt sander and the trajectory of the robot are adaptively matched to the workpiece surface, overcoming the problems of over-grinding, under-grinding and uneven surface quality caused by the mismatch between pressure and speed in grinding complex curved surfaces.

[0007] This invention is implemented as follows: a collaborative control method for belt sanders in grinding and polishing robots, applied to communication collaborative equipment, specifically includes the following steps: S11: Establish a real-time communication network based on EtherCAT and Modbus to connect and communicate with robots, belt sanders, PLC controllers, multi-dimensional force sensors and 3D vision modules, realize data synchronization and command alignment, and generate a unified underlying data platform. S12: Based on the underlying data platform, the real-time communication network identifies the surface features of the workpiece through the 3D vision module and calls the process database to pre-generate the robot motion trajectory and the speed change curve of the belt sander that match the curvature of the surface, so as to realize the initial coordination of trajectory planning and speed setting. S13: The multi-dimensional force sensor collects contact force data in real time during the execution of the planned trajectory. Through the force-position hybrid control algorithm, it synchronously and dynamically adjusts the running speed of the belt sander and the feeding posture of the robot to compensate for trajectory errors with a constant output contact force. S14: Through the coordinated control of the PLC controller, the belt sander and the robot can quickly switch between the upper and lower double material tables, accurately realize the scheduling of work stations for continuous production, complete the alternating connection of grinding and material preparation, eliminate equipment waiting, and build an uninterrupted continuous production cycle. S15: Establish a fault monitoring and linkage safety response module in the underlying data platform. In the event of a fault, immediately coordinate the shutdown and record the operating parameters. At the same time, continuously store process data into the process library and iteratively optimize the coordination parameters through a self-learning algorithm.

[0008] Furthermore, in S11, a real-time communication network based on EtherCAT and Modbus is established to interconnect the robot, belt sander, PLC controller, multi-dimensional force sensor, and 3D vision module, including: Using EtherCAT industrial Ethernet as the core backbone network, it directly connects to the grinding and polishing robot controller, sander frequency converter and PLC controller to complete the deployment of a layered real-time communication architecture. The EtherCAT master station is integrated into the robot controller. It uses the Modbus-TCP protocol as an auxiliary communication channel through the standard RJ45 interface and shielded twisted pair cable. Through the embedded gateway, the multi-dimensional force sensor and 3D vision module are seamlessly connected to the same EtherCAT network domain. The system time of all nodes is precisely synchronized using the EtherCAT industrial Ethernet distributed clock mechanism, and a synchronization manager is configured for each slave station to ensure that the robot's trajectory instructions, the start / stop speed adjustment commands of the belt sander, and the PLC's station control signals are executed within a strictly aligned timing window.

[0009] Furthermore, data synchronization and command alignment are achieved to generate a unified underlying data platform, including: For multi-dimensional force sensors and 3D vision modules, a unified data mapping table is defined to standardize the output torque signals and point cloud coordinate data into a fixed frame format that includes a data header, payload, and check code. By configuring Modbus-TCP communication parameters and setting a millisecond-level transmission cycle, key sensing data can be uploaded to the EtherCAT master station in real time and cyclically. The EtherCAT master station uses a precise distributed clock to assign a unified timestamp to each received frame of data, completing the timestamp alignment of cross-domain heterogeneous data. Then, it calls the data fusion engine to perform spatial correlation and logical verification of force information and visual coordinates, eliminate outliers, and integrate the real-time status of the robot and belt sander to form a globally consistent and time-synchronized panoramic view of the equipment status.

[0010] Furthermore, in S12, based on the real-time communication network of the underlying data platform, the 3D vision module identifies the surface features of the workpiece and calls the process database to pre-generate the robot motion trajectory and the speed change curve of the belt sander that match the curvature of the surface, including: Based on the underlying data platform, the 3D vision module first acquires the workpiece point cloud data with high precision and calculates the curvature value of each point on its surface in real time. Then, it calls the process database to match the current workpiece material and target roughness process requirements with the real-time curvature distribution map, and pre-generates a theoretically optimal sander linear speed reference value for each point on the trajectory. During the grinding process, the actual curvature of the current contact point is mapped based on the real-time pose of the robot end effector. By dynamically matching the convex or concave areas with large curvature, the speed of the belt sander is automatically reduced to reduce the amount of material removed per unit time and prevent over-grinding. In flat or low-curvature areas, the belt sander speed is automatically increased to ensure grinding efficiency. The speed is dynamically adjusted based on the real-time curvature, which effectively avoids problems such as over-grinding, missed grinding, or uneven surface quality caused by abrupt changes in curvature.

[0011] Furthermore, in S13, the running speed of the belt sander and the feed posture of the robot are synchronously and dynamically adjusted through a force-position hybrid control algorithm to compensate for trajectory errors with a constant output contact force, including: The force-position hybrid control algorithm receives and processes feedback from multi-dimensional force sensors in real time, calculates the deviation between the current contact force and the set threshold, and realizes the first level of the adjustment process: dynamically adjust the running speed of the belt sander. When the real-time contact force exceeds the set upper limit, the belt sander linear speed is increased immediately to reduce the cutting depth per unit time. Conversely, the speed is reduced to increase the cutting action. The contact force is initially stabilized by rapidly changing the cutting conditions. While adjusting the speed, the force-position hybrid control algorithm initiates the second level of compensation: synchronously fine-tuning the robot's normal posture to compensate for trajectory deviations caused by workpiece clamping errors, robot absolute positioning errors, or workpiece deformation in real time. The speed and position adjustments are deeply coupled and jointly output to ensure that the contact pressure between the abrasive belt and the workpiece remains stable within a constant range, achieving high-precision force control consistency during dynamic grinding.

[0012] Furthermore, the force-position hybrid control algorithm dynamically adjusts the operating speed of the belt sander and fine-tunes the normal posture of the robot. The algorithm includes: The speed adjustment formula of the force-position hybrid control algorithm is as follows: ; Among them, V b (t) represents the final commanded linear velocity of the belt sander at time t; V b0 (t) represents the baseline velocity generated by trajectory-velocity co-planning; ΔF(t) = F actual (t)-F desired The deviation between the real-time contact force and the set desired force; K p1 K i1 These are the proportional and integral control gains of the speed loop, respectively. The attitude compensation formula of the force-position hybrid control algorithm is as follows: ; ; Where ΔP(t) is the position compensation amount of the robot end effector along the feed direction at time t; This is the normal vector attitude compensation amount in the tool coordinate system; K is the vector deviation between the real-time contact torque and the desired torque. p2 K i2 K p3 K i3 The ratio and integral gain of the corresponding position and attitude compensation loop; Comprehensive adjustment of the output formula: ;

[0013] Multi-loop PID calculations are performed using the parameter matrix K, ultimately outputting the adjustments to speed, position, and attitude synchronously. In the above formula: t is the current time of the real-time control system, which is incremented in units of control cycle; V b (t) represents the commanded linear velocity of the belt sander; V b0 (t) represents the baseline velocity of the belt sander; F actual(t) represents the real-time contact force measurement; F desired Set the desired contact force value; ΔF(t) represents the real-time contact force deviation; K p1 K i1为 Speed ​​loop PID control gain; ΔP(t) is the robot position compensation amount, which is the displacement adjustment amount of the robot end effector along the feed direction output by the algorithm; This is the robot's posture compensation amount; This refers to the real-time contact torque deviation. K p2 K i2 K p3 K i3 The PID control gain for the position and attitude compensation loop corresponds to the proportional and integral coefficients of the position compensation and attitude compensation control loops, respectively, to achieve decoupling and fine compensation for different degrees of freedom. U(t) is the integrated control output vector, which is the set of adjustment instructions for all actuators calculated synchronously at time t; f PID (·) is a multivariable PID control law function, representing the process of performing multi-loop parallel PID calculations with torque deviation as input and gain matrix K as parameter; K is the global PID gain matrix, containing all independent PID gain parameters (K0). p1 K i1 ...) matrix; T c To control the cycle, the time it takes for the algorithm to execute one loop is synchronized with the communication network cycle.

[0014] Furthermore, in S14, the PLC controller coordinates the rapid switching between the belt sander and the robot between the upper and lower material tables, precisely achieving continuous production scheduling at the workstations, including: After receiving the upper station completion signal from the robot, the PLC controller immediately sends a sequence of instructions to the frequency converter of the belt sander. The sequence of instructions controls the belt sander to perform precise deceleration, short pauses and station switching actions, driving its working contact wheel to move quickly and smoothly and accurately align with the workpiece to be processed on the lower platform. At the same time, it instructs the lower platform to make fine adjustments to ensure that the workpiece is in the best grinding posture. While the belt sander is switching stations, the PLC controller simultaneously sends instructions to the AGV or conveyor mechanism associated with the upper material platform, controlling it to move the finished workpiece out of the work area and immediately replenish the new blank to the upper material platform. The switching process, including equipment relocation and material flow, achieves millisecond-level precise connection between the belt sander and robot tasks, completely eliminating equipment waiting time.

[0015] Furthermore, in S15, a fault monitoring and coordinated safety response module is established in the underlying data platform. In the event of a fault, this module immediately coordinates a shutdown and records operating parameters, including: A distributed fault monitoring network was built in the underlying data platform to poll and collect the current, temperature, and communication status of the belt sander, the joint position, torque and error code of the robot, and the health status of each sensor and actuator in real time. The underlying data platform performs fast filtering and feature extraction on the collected raw data and compares it with the preset safety threshold in real time. Once any parameter exceeds the allowable range or a signal interruption occurs, a high-priority fault flag is immediately triggered. Once the fault flag is triggered, the linkage safety response module immediately activates the preset collaborative shutdown sequence. The collaborative shutdown sequence forces the belt sander to perform emergency braking, instructs the robot to pause its current trajectory and return to the safe standby point at a constant speed, and simultaneously locks the actions of all material platforms and transmission mechanisms. During the shutdown process, the linkage safety response module synchronously captures and stores snapshots of the entire system's operating parameters within the critical time window before and after the fault occurs, including real-time force values, speed, position, and equipment status codes, providing a complete data link for post-event analysis and process optimization.

[0016] Compared with the prior art, the present invention provides a collaborative control system and method for belt sanders for grinding and polishing robots, which has the following advantages: 1. Through the full-dimensional collaborative control of trajectory, speed, force control, workstation, and safety, the problems of traditional grinding and polishing automation are effectively solved. The dynamic collaborative mechanism of force-position coupling realizes the adaptive matching of the sander speed and robot trajectory to the workpiece surface, overcoming the problems of over-grinding, under-grinding, and uneven surface quality caused by the mismatch between pressure and speed in grinding complex curved surfaces, and significantly improving the consistency and accuracy of processing. In addition, the seamless alternation collaboration mechanism of dual workstations and robots eliminates the equipment idle waiting caused by loading and unloading through the PLC controller, breaking through the efficiency bottleneck of single workstation production. At the same time, the safety-process collaborative control of fault linkage transforms passive shutdown into an active data learning process, and uses fault conditions to back-optimize process parameters, enhancing the reliability and process robustness of the system. 2. By introducing collaborative parameter self-optimization and digital twin collaborative simulation, it possesses powerful self-learning and pre-verification capabilities, which can shorten the process debugging time for new workpieces and new materials by more than 90% and virtually eliminate interference risks, greatly improving the response speed and first-piece success rate of flexible production changeover. The expansion of multi-sand belt machine cluster collaborative control and AGV-sand belt machine-robot full-process collaborative closed loop upgrades from single-machine collaboration to full-workshop-level automated production flow, realizing seamless integration of processes from rough grinding to polishing and automatic material flow, and promoting the evolution of production lines towards unmanned and intelligent directions.

[0017] A collaborative control system for a belt sander for a grinding and polishing robot, used to execute the above-mentioned collaborative control method, the collaborative control system comprising: The communication and collaboration module is used to establish and maintain real-time data links and platforms between robots, belt sanders, PLCs, sensors, and 3D vision components; The trajectory and speed planning module is used to pre-generate a robot trajectory and sander speed coordination scheme based on visual recognition and process data; The force-position coupling control module is used to process force feedback in real time and adjust the speed of the belt sander and the robot posture in conjunction to maintain constant force output; The dual-station scheduling module is used to coordinate the control of the belt sander, robot and loading / unloading mechanism to achieve seamless switching of stations and continuous production; The safety and process optimization module is used to monitor system faults, execute collaborative safety strategies, and iteratively optimize collaborative parameters based on process data.

[0018] Specifically, the force-position coupling control module includes: The signal acquisition and processing unit is used to receive and filter the contact force and torque data from the six-dimensional force sensor in real time. Multi-loop calculation unit is used to run the speed adjustment and pose compensation algorithm in parallel, and to calculate the speed adjustment of the belt sander and the posture compensation of the robot. The collaborative instruction output unit is used to superimpose the calculated adjustment amount onto the reference instruction and simultaneously send it to the belt sander and robot controller. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a collaborative control method for belt sanders for grinding and polishing robots proposed in this invention. Figure 2 This is a schematic diagram of the structure of a belt sander collaborative control system for a grinding and polishing robot proposed in this invention; Figure 3 This is a schematic diagram of the force-position coupling control module in a belt sander collaborative control system for a grinding and polishing robot proposed in this invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0023] The implementation of the present invention will be described in detail below with reference to specific embodiments.

[0024] In the accompanying drawings of this embodiment, the same or similar reference numerals correspond to the same or similar components. In the description of this invention, it should be understood that if terms such as "upper," "lower," "left," and "right" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting this invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0025] Reference Figure 1 As shown, a collaborative control method for belt sanders in a grinding and polishing robot, applied to a communication collaborative device, specifically includes the following steps: S11: Establish a real-time communication network based on EtherCAT and Modbus to connect and communicate with robots, belt sanders, PLC controllers, multi-dimensional force sensors and 3D vision modules, realize data synchronization and command alignment, and generate a unified underlying data platform. This platform has data caching and priority management functions to ensure the real-time transmission of key control commands under high load and to provide a unified application programming interface for upper-layer applications. This includes establishing a real-time communication network based on EtherCAT and Modbus to interconnect the robot, belt sander, PLC controller, multi-dimensional force sensor, and 3D vision module, including: Using EtherCAT industrial Ethernet as the core backbone network, it directly connects to the grinding and polishing robot controller, sander frequency converter and PLC controller to complete the deployment of a layered real-time communication architecture. The EtherCAT master station is integrated into the robot controller. It uses the Modbus-TCP protocol as an auxiliary communication channel through the standard RJ45 interface and shielded twisted pair cable. Through the embedded gateway, the multi-dimensional force sensor and 3D vision module are seamlessly connected to the same EtherCAT network domain. The system time of all nodes is precisely synchronized using the EtherCAT industrial Ethernet distributed clock mechanism, and a synchronization manager is configured for each slave station to ensure that the robot's trajectory instructions, the start / stop speed adjustment commands of the belt sander, and the PLC's station control signals are executed within a strictly aligned timing window. S12: Based on the underlying data platform, the real-time communication network identifies the surface features of the workpiece through the 3D vision module and calls the process database to pre-generate the robot motion trajectory and the speed change curve of the belt sander that match the curvature of the surface, realizing the initial coordination of trajectory planning and speed setting; this process includes real-time noise reduction and feature enhancement processing of workpiece point cloud data, and verifies the feasibility of trajectory and speed curves through simulation engine to avoid potential motion interference. S13: The multi-dimensional force sensor collects contact force data in real time during the execution of the planned trajectory. The force-position hybrid control algorithm synchronously and dynamically adjusts the running speed of the belt sander and the feeding posture of the robot to compensate for trajectory errors with a constant output contact force. The algorithm has a built-in feedforward compensation mechanism that can predict the force change trend based on the curvature and fine-tune the parameters in advance to suppress force overshoot and ensure the smoothness and stability of the contact force response. Among them, the force-position hybrid control algorithm synchronously and dynamically adjusts the sander's running speed and the robot's feed posture to compensate for trajectory errors with a constant output contact force, including: The force-position hybrid control algorithm receives and processes feedback from multi-dimensional force sensors in real time, calculates the deviation between the current contact force and the set threshold, and realizes the first level of the adjustment process: dynamically adjust the running speed of the belt sander. When the real-time contact force exceeds the set upper limit, the belt sander linear speed is increased immediately to reduce the cutting depth per unit time. Conversely, the speed is reduced to increase the cutting action. The contact force is initially stabilized by rapidly changing the cutting conditions. While adjusting the speed, the force-position hybrid control algorithm initiates the second level of compensation: synchronously fine-tuning the robot's normal posture, and real-time compensation for trajectory deviations caused by workpiece clamping errors, robot absolute positioning errors, or workpiece deformation. The speed and position adjustments are deeply coupled and jointly output to ensure that the contact pressure between the sanding belt and the workpiece remains stable within a constant range, achieving high-precision force control consistency during dynamic grinding. S14: Through the coordinated control of the PLC controller, the belt sander and the robot can quickly switch between the upper and lower double material tables, accurately realize the scheduling of continuous production of the work station, complete the alternation of grinding and material preparation, eliminate equipment waiting, and build an uninterrupted continuous production cycle; the scheduling logic includes equipment status interlocking and sequence control to ensure that in the switching action chain, the safe completion of the previous process is a necessary condition for triggering the next process, while maintaining the dynamic balance of the production cycle; S15: Establish a fault monitoring and coordinated safety response module in the underlying data platform. Upon a fault, immediately coordinate a shutdown and record operating parameters. Simultaneously, continuously store process data in the process library and iteratively optimize the coordinated parameters through a self-learning algorithm. This module automatically classifies and correlates the recorded operating parameters to pinpoint the fault causes and provides weighted optimization directions for the self-learning algorithm, thereby improving the targeting and efficiency of parameter iteration. Specifically, a fault monitoring and coordinated safety response module is established in the underlying data platform. In the event of a fault, it immediately coordinates a shutdown and records operating parameters, including: A distributed fault monitoring network was built in the underlying data platform to poll and collect the current, temperature, and communication status of the belt sander, the joint position, torque and error code of the robot, and the health status of each sensor and actuator in real time. The underlying data platform performs fast filtering and feature extraction on the collected raw data and compares it with the preset safety threshold in real time. Once any parameter exceeds the allowable range or a signal interruption occurs, a high-priority fault flag is immediately triggered. Once the fault flag is triggered, the linkage safety response module immediately activates the preset collaborative shutdown sequence. The collaborative shutdown sequence forces the belt sander to perform emergency braking, instructs the robot to pause its current trajectory and return to the safe standby point at a constant speed, and simultaneously locks the actions of all material platforms and transmission mechanisms. During the shutdown process, the linkage safety response module synchronously captures and stores snapshots of the entire system's operating parameters within the critical time window before and after the fault occurs, including real-time force values, speed, position, and equipment status codes, providing a complete data link for post-event analysis and process optimization.

[0026] In S11 of this embodiment, data synchronization and instruction alignment are achieved to generate a unified underlying data platform, including: For multi-dimensional force sensors and 3D vision modules, a unified data mapping table is defined to standardize the output torque signals and point cloud coordinate data into a fixed frame format that includes a data header, payload, and check code. By configuring Modbus-TCP communication parameters and setting a millisecond-level transmission cycle, key sensing data can be uploaded to the EtherCAT master station in real time and cyclically. The EtherCAT master station uses a precise distributed clock to assign a unified timestamp to each received frame of data, completing the timestamp alignment of cross-domain heterogeneous data. Then, it calls the data fusion engine to perform spatial correlation and logical verification of force information and visual coordinates, eliminate outliers, and integrate the real-time status of the robot and belt sander to form a globally consistent and time-synchronized panoramic view of the equipment status.

[0027] In S12 of this embodiment, based on the real-time communication network of the underlying data platform, the 3D vision module identifies the surface features of the workpiece and calls the process database to pre-generate a robot motion trajectory and a sander speed change curve that match the surface curvature, including: Based on the underlying data platform, the 3D vision module first acquires the workpiece point cloud data with high precision and calculates the curvature value of each point on its surface in real time. Then, it calls the process database to match the current workpiece material and target roughness process requirements with the real-time curvature distribution map, and pre-generates a theoretically optimal sander linear speed reference value for each point on the trajectory. During the grinding process, the actual curvature of the current contact point is mapped based on the real-time pose of the robot end effector. By dynamically matching the convex or concave areas with large curvature, the speed of the belt sander is automatically reduced to reduce the amount of material removed per unit time and prevent over-grinding. In flat or low-curvature areas, the belt sander speed is automatically increased to ensure grinding efficiency. The speed is dynamically adjusted based on the real-time curvature, which effectively avoids problems such as over-grinding, missed grinding, or uneven surface quality caused by abrupt changes in curvature.

[0028] In S13 of this embodiment, the force-position hybrid control algorithm dynamically adjusts the running speed of the belt sander and fine-tunes the normal orientation of the robot. The algorithm includes: Speed ​​adjustment formula for force-position hybrid control algorithm: ;

[0029] Among them, V b (t) represents the final commanded linear velocity of the belt sander at time t; V b0 (t) represents the baseline velocity generated by trajectory-velocity co-planning; ΔF(t) = F actual (t)-F desired The deviation between the real-time contact force and the set desired force; K p1 K i1 These are the proportional and integral control gains of the speed loop, respectively. The attitude compensation formula for the force-position hybrid control algorithm is as follows: ; ;

[0030] Where ΔP(t) is the position compensation amount of the robot end effector along the feed direction at time t; This is the normal vector attitude compensation amount in the tool coordinate system; K is the vector deviation between the real-time contact torque and the desired torque. p2 K i2 K p3 K i3The ratio and integral gain of the corresponding position and attitude compensation loop; Comprehensive adjustment of the output formula: ;

[0031] Multi-loop PID calculations are performed using the parameter matrix K, ultimately outputting the adjustments to speed, position, and attitude synchronously. In the above formula: t is the current time of the real-time control system, which is incremented in units of control cycle; V b (t) represents the commanded linear velocity of the belt sander, which is dynamically adjusted according to the force deviation to achieve rapid adjustment of the cutting action; V b0 (t) represents the baseline speed of the belt sander, which is a preset value obtained from S12 to achieve initial coordination between trajectory planning and speed setting, and serves as the reference point for speed adjustment. F actual (t) is the real-time contact force measurement value, which is a normal contact force scalar acquired by a multi-dimensional force sensor at time t and filtered, in N; F desired The desired contact force is set by retrieving the optimal process parameters from the process database based on the current workpiece material and abrasive belt type, which can be either a constant threshold or a segmented threshold. ΔF(t) is the real-time contact force deviation, which is the main input of the control algorithm. A positive value indicates overpressure, requiring deceleration or retraction; a negative value indicates underpressure, requiring acceleration or engagement. K p1 K i1 For the speed loop PID control gain, identify the proportional and integral coefficients to be tuned, which are used to map the force deviation into the adjustment amount of the belt sander speed. The integral term is used to eliminate steady-state error. ΔP(t) is the robot position compensation amount, which is the displacement adjustment amount of the robot end effector along the feed direction output by the algorithm; The tool coordinate system normal vector output by the algorithm is used to adjust the tool posture in real time to ensure the optimal contact angle for robot posture compensation. For real-time contact torque deviation, the deviation between the torque vector measured by a multi-dimensional force sensor and the desired torque is used to detect and compensate for the torsion or tilt of the tool. K p2 K i2 K p3 K i3 The PID control gain for the position and attitude compensation loop corresponds to the proportional and integral coefficients of the position compensation and attitude compensation control loops, respectively, to achieve decoupling and fine compensation for different degrees of freedom. U(t) is the integrated control output vector, which is the set of adjustment instructions for all actuators calculated synchronously at time t; f PID (·) is a multivariable PID control law function, representing the process of performing multi-loop parallel PID calculations with torque deviation as input and gain matrix K as parameter; K is the global PID gain matrix, containing all independent PID gain parameters (K0). p1 K i1 ...) matrix; T c To control the cycle, the time it takes for the algorithm to execute one loop is synchronized with the communication network cycle.

[0032] In S14 of this embodiment, the PLC controller coordinates the rapid switching between the belt sander and the robot between the upper and lower double material tables, precisely achieving continuous production scheduling at the workstations, including: After receiving the upper station completion signal from the robot, the PLC controller immediately sends a sequence of instructions to the frequency converter of the belt sander. The sequence of instructions controls the belt sander to perform precise deceleration, short pauses and station switching actions, driving its working contact wheel to move quickly and smoothly and accurately align with the workpiece to be processed on the lower platform. At the same time, it instructs the lower platform to make fine adjustments to ensure that the workpiece is in the best grinding posture. While the belt sander is switching stations, the PLC controller simultaneously sends instructions to the AGV or conveyor mechanism associated with the upper material platform, controlling it to move the finished workpiece out of the work area and immediately replenish the new blank to the upper material platform. The switching process, including equipment relocation and material flow, achieves millisecond-level precise connection between the belt sander and robot tasks, completely eliminating equipment waiting time.

[0033] This technical solution effectively solves the problems of traditional grinding and polishing automation through full-dimensional collaborative control of trajectory, speed, force control, workstation, and safety. The dynamic collaborative mechanism of force-position coupling achieves adaptive matching between the speed of the belt sander and the robot trajectory on the workpiece surface, overcoming the problems of over-grinding, under-grinding, and uneven surface quality caused by pressure and speed mismatch in grinding complex curved surfaces. This significantly improves processing consistency and accuracy. Furthermore, the seamless alternating collaborative mechanism between dual workstations and the robot, through the PLC controller, eliminates equipment idle waiting caused by loading and unloading, breaking through the efficiency bottleneck of single-workstation production. Simultaneously, the fault-linked safety-process collaborative control transforms passive shutdown into an active data learning process, using fault conditions to reverse-optimize process parameters, enhancing the system's reliability and process robustness.

[0034] Reference Figure 2-3As shown, a collaborative control system for a belt sander for a grinding and polishing robot is used to execute the aforementioned collaborative control method. The collaborative control system includes: a communication collaboration module for establishing and maintaining real-time data links and platforms between the robot, belt sander, PLC, sensors, and 3D vision components; a trajectory and speed planning module for pre-generating a robot trajectory and belt sander speed collaboration scheme based on visual recognition and process data; a force-position coupling control module for real-time processing of force feedback and linkage adjustment of belt sander speed and robot posture to maintain constant force output; a dual-station scheduling module for collaborative control of the belt sander, robot, and loading / unloading mechanisms to achieve seamless station switching and continuous production; and a safety and process optimization module for monitoring system faults, executing collaborative safety strategies, and iteratively optimizing collaborative parameters based on process data. The seamless alternation collaboration mechanism between the dual-station and robot eliminates equipment idle waiting caused by loading / unloading through the PLC controller, breaking through the efficiency bottleneck of single-station production. At the same time, the fault-linked safety-process collaborative control transforms passive shutdown into an active data learning process, using fault conditions to reverse-optimize process parameters, thereby enhancing the system's reliability and process robustness.

[0035] In this embodiment, the force-position coupling control module includes: a signal acquisition and processing unit for receiving and filtering contact force and torque data from a six-dimensional force sensor in real time; a multi-loop calculation unit for running speed adjustment and pose compensation algorithms in parallel to calculate the speed adjustment of the belt sander and the posture compensation of the robot; and a collaborative command output unit for superimposing the calculated adjustment amount onto the reference command and simultaneously sending it to the belt sander and robot controller. By introducing collaborative parameter self-optimization and digital twin collaborative simulation, it possesses powerful self-learning and pre-verification capabilities, which can shorten the process debugging time for new workpieces and new materials by more than 90%, and virtually eliminate interference risks, greatly improving the response speed and first-piece success rate of flexible production changeover. This technical solution expands the multi-sand belt machine cluster collaborative control and the AGV-sand belt machine-robot full-process collaborative closed loop, upgrading from single-machine collaboration to full-workshop-level automated production flow. It achieves seamless integration of processes from rough grinding to polishing and automatic material flow, and promotes the evolution of production lines towards unmanned and intelligent directions.

[0036] In this embodiment, the entire operation process can be automated by computer control. In each operation stage, sensors can be set up to provide signal feedback and ensure that the steps are performed sequentially. These are all conventional knowledge of current automation control, and will not be elaborated on in this embodiment.

[0037] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A collaborative control method for belt sanders for grinding and polishing robots, characterized in that, Applied to communication collaboration devices, the specific steps include: S11: Establish a real-time communication network based on EtherCAT and Modbus to connect and communicate with robots, belt sanders, PLC controllers, multi-dimensional force sensors and 3D vision modules, realize data synchronization and command alignment, and generate a unified underlying data platform. S12: Based on the underlying data platform, the real-time communication network identifies the surface features of the workpiece through the 3D vision module and calls the process database to pre-generate the robot motion trajectory and the speed change curve of the belt sander that match the curvature of the surface, so as to realize the initial coordination of trajectory planning and speed setting. S13: The multi-dimensional force sensor collects contact force data in real time during the execution of the planned trajectory. Through the force-position hybrid control algorithm, it synchronously and dynamically adjusts the running speed of the belt sander and the feeding posture of the robot to compensate for trajectory errors with a constant output contact force. S14: Through the coordinated control of the PLC controller, the belt sander and the robot can quickly switch between the upper and lower double material tables, accurately realize the scheduling of work stations for continuous production, complete the alternating connection of grinding and material preparation, eliminate equipment waiting, and build an uninterrupted continuous production cycle. S15: Establish a fault monitoring and linkage safety response module in the underlying data platform. In the event of a fault, immediately coordinate the shutdown and record the operating parameters. At the same time, continuously store process data into the process library and iteratively optimize the coordination parameters through a self-learning algorithm.

2. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 1, characterized in that, In S11, a real-time communication network based on EtherCAT and Modbus is established to interconnect the robot, belt sander, PLC controller, multi-dimensional force sensor, and 3D vision module, including: Using EtherCAT industrial Ethernet as the core backbone network, it directly connects to the grinding and polishing robot controller, sander frequency converter and PLC controller to complete the deployment of a layered real-time communication architecture. The EtherCAT master station is integrated into the robot controller. It uses the Modbus-TCP protocol as an auxiliary communication channel through the standard RJ45 interface and shielded twisted pair cable. Through the embedded gateway, the multi-dimensional force sensor and 3D vision module are seamlessly connected to the same EtherCAT network domain. The system time of all nodes is precisely synchronized using the EtherCAT industrial Ethernet distributed clock mechanism, and a synchronization manager is configured for each slave station to ensure that the robot's trajectory instructions, the start / stop speed adjustment commands of the belt sander, and the PLC's station control signals are executed within a strictly aligned timing window.

3. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 2, characterized in that, To achieve data synchronization and command alignment, and generate a unified underlying data platform, including: For multi-dimensional force sensors and 3D vision modules, a unified data mapping table is defined to standardize the output torque signals and point cloud coordinate data into a fixed frame format that includes a data header, payload, and check code. By configuring Modbus-TCP communication parameters and setting a millisecond-level transmission cycle, key sensing data can be uploaded to the EtherCAT master station in real time and cyclically. The EtherCAT master station uses a precise distributed clock to assign a unified timestamp to each received frame of data, completing the timestamp alignment of cross-domain heterogeneous data. Then, it calls the data fusion engine to perform spatial correlation and logical verification of force information and visual coordinates, eliminate outliers, and integrate the real-time status of the robot and belt sander to form a globally consistent and time-synchronized panoramic view of the equipment status.

4. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 3, characterized in that, In S12, based on the real-time communication network of the underlying data platform, the 3D vision module identifies the surface features of the workpiece and calls the process database to pre-generate robot motion trajectories and belt sander speed change curves that match the surface curvature, including: Based on the underlying data platform, the 3D vision module first acquires the workpiece point cloud data with high precision and calculates the curvature value of each point on its surface in real time. Then, it calls the process database to match the current workpiece material and target roughness process requirements with the real-time curvature distribution map, and pre-generates a theoretically optimal sander linear speed reference value for each point on the trajectory. During the grinding process, the actual curvature of the current contact point is mapped based on the real-time pose of the robot end effector. By dynamically matching the convex or concave areas with large curvature, the speed of the belt sander is automatically reduced to reduce the amount of material removed per unit time and prevent over-grinding. In flat or low-curvature areas, the belt sander speed is automatically increased to ensure grinding efficiency. The speed is dynamically adjusted based on the real-time curvature, which effectively avoids problems such as over-grinding, missed grinding, or uneven surface quality caused by abrupt changes in curvature.

5. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 4, characterized in that, In S13, the running speed of the belt sander and the robot's feed posture are synchronously and dynamically adjusted through a force-position hybrid control algorithm to compensate for trajectory errors with a constant output contact force, including: The force-position hybrid control algorithm receives and processes feedback from multi-dimensional force sensors in real time, calculates the deviation between the current contact force and the set threshold, and realizes the first level of the adjustment process: dynamically adjust the running speed of the belt sander. When the real-time contact force exceeds the set upper limit, the belt sander linear speed is increased immediately to reduce the cutting depth per unit time. Conversely, the speed is reduced to increase the cutting action. The contact force is initially stabilized by rapidly changing the cutting conditions. While adjusting the speed, the force-position hybrid control algorithm initiates the second level of compensation: synchronously fine-tuning the robot's normal posture to compensate for trajectory deviations caused by workpiece clamping errors, robot absolute positioning errors, or workpiece deformation in real time. The speed and position adjustments are deeply coupled and jointly output to ensure that the contact pressure between the abrasive belt and the workpiece remains stable within a constant range, achieving high-precision force control consistency during dynamic grinding.

6. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 5, characterized in that, The force-position hybrid control algorithm dynamically adjusts the running speed of the belt sander and fine-tunes the normal posture of the robot. The algorithm includes: The speed adjustment formula of the force-position hybrid control algorithm is as follows: ； Among them, V b (t) represents the final commanded linear velocity of the belt sander at time t; V b0 (t) represents the baseline velocity generated by trajectory-velocity co-planning; ΔF(t) = F actual (t)-F desired The deviation between the real-time contact force and the set desired force; K p1 K i1 These are the proportional and integral control gains of the speed loop, respectively. The attitude compensation formula of the force-position hybrid control algorithm is as follows: ； ； Where ΔP(t) is the position compensation amount of the robot end effector along the feed direction at time t; This is the normal vector attitude compensation amount in the tool coordinate system; K is the vector deviation between the real-time contact torque and the desired torque. p2 K i2 K p3 K i3 The ratio and integral gain of the corresponding position and attitude compensation loop; Comprehensive adjustment of the output formula: ； Multi-loop PID calculations are performed using the parameter matrix K, ultimately outputting the adjustments to speed, position, and attitude synchronously. In the above formula: t is the current time of the real-time control system, which is incremented in units of control cycle; V b (t) represents the commanded linear velocity of the belt sander; V b0 (t) represents the baseline velocity of the belt sander; F actual (t) represents the real-time contact force measurement; F desired Set the desired contact force value; ΔF(t) represents the real-time contact force deviation; K p1 K i1为 Speed ​​loop PID control gain; ΔP(t) is the robot position compensation amount, which is the displacement adjustment amount of the robot end effector along the feed direction output by the algorithm; This is the robot's posture compensation amount; For real-time contact torque deviation; K p2 K i2 K p3 K i3 The PID control gain for the position and attitude compensation loop corresponds to the proportional and integral coefficients of the position compensation and attitude compensation control loops, respectively, to achieve decoupling and fine compensation for different degrees of freedom. U(t) is the integrated control output vector, which is the set of adjustment instructions for all actuators calculated synchronously at time t; f PID (·) is a multivariable PID control law function, representing the process of performing multi-loop parallel PID calculations with torque deviation as input and gain matrix K as parameter; K is the global PID gain matrix, containing all independent PID gain parameters (K0). p1 K i1 ...) matrix; T c To control the cycle, the time it takes for the algorithm to execute one loop is synchronized with the communication network cycle.

7. A method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 6, characterized in that, In S14, the PLC controller coordinates the rapid switching between the belt sander and the robot between the upper and lower material tables, precisely achieving continuous production scheduling at the workstations, including: After receiving the upper station completion signal from the robot, the PLC controller immediately sends a sequence of instructions to the frequency converter of the belt sander. The sequence of instructions controls the belt sander to perform precise deceleration, short pauses and station switching actions, driving its working contact wheel to move quickly and smoothly and accurately align with the workpiece to be processed on the lower platform. At the same time, it instructs the lower platform to make fine adjustments to ensure that the workpiece is in the best grinding posture. While the belt sander is switching stations, the PLC controller simultaneously sends instructions to the AGV or conveyor mechanism associated with the upper material platform, controlling it to move the finished workpiece out of the work area and immediately replenish the new blank to the upper material platform. The switching process, including equipment relocation and material flow, achieves millisecond-level precise connection between the belt sander and robot tasks, completely eliminating equipment waiting time.

8. The method for collaborative control of a belt sander for a grinding and polishing robot as described in claim 7, characterized in that, In S15, a fault monitoring and coordinated safety response module is established in the underlying data platform. In the event of a fault, it immediately coordinates a shutdown and records operating parameters, including: A distributed fault monitoring network was built in the underlying data platform to poll and collect the current, temperature, and communication status of the belt sander, the joint position, torque and error code of the robot, and the health status of each sensor and actuator in real time. The underlying data platform performs fast filtering and feature extraction on the collected raw data and compares it with the preset safety threshold in real time. Once any parameter exceeds the allowable range or a signal interruption occurs, a high-priority fault flag is immediately triggered. Once the fault flag is triggered, the linkage safety response module immediately activates the preset collaborative shutdown sequence. The collaborative shutdown sequence forces the belt sander to perform emergency braking, instructs the robot to pause its current trajectory and return to the safe standby point at a constant speed, and simultaneously locks the actions of all material platforms and transmission mechanisms. During the shutdown process, the linkage safety response module synchronously captures and stores snapshots of the entire system's operating parameters within the critical time window before and after the fault occurs, including real-time force values, speed, position, and equipment status codes, providing a complete data link for post-event analysis and process optimization.

9. A collaborative control system for a belt sander for a grinding and polishing robot, characterized in that, The cooperative control system is used to perform the cooperative control method according to any one of claims 1-8, wherein the cooperative control system comprises: The communication and collaboration module is used to establish and maintain real-time data links and platforms between robots, belt sanders, PLCs, sensors, and 3D vision components; The trajectory and speed planning module is used to pre-generate a robot trajectory and sander speed coordination scheme based on visual recognition and process data; The force-position coupling control module is used to process force feedback in real time and adjust the speed of the belt sander and the robot posture in conjunction to maintain constant force output; The dual-station scheduling module is used to coordinate the control of the belt sander, robot and loading / unloading mechanism to achieve seamless switching of stations and continuous production; The safety and process optimization module is used to monitor system faults, execute collaborative safety strategies, and iteratively optimize collaborative parameters based on process data.

10. A collaborative control system for a belt sander for a grinding and polishing robot as described in claim 9, characterized in that, The force-position coupling control module includes: The signal acquisition and processing unit is used to receive and filter the contact force and torque data from the six-dimensional force sensor in real time. Multi-loop calculation unit is used to run the speed adjustment and pose compensation algorithm in parallel, and to calculate the speed adjustment of the belt sander and the posture compensation of the robot. The collaborative instruction output unit is used to superimpose the calculated adjustment amount onto the reference instruction and simultaneously send it to the belt sander and robot controller.

Images