Automatic equipment control system and method and storage medium

The surface model of the target polished parts is constructed through high-frequency laser scanner and B-spline surface fitting technology, and the six-dimensional torque sensor and vibration sensor are combined for real-time contact status monitoring to achieve reverse impedance control, solving the problem of grinding force fluctuations in traditional polishing technology, significantly improving polishing accuracy and production efficiency.

CN120122558APending Publication Date: 2025-06-10SHANGHAI DAOCHENG MACHINERY EQUIPMENT CO LTD
View PDF 0 Cites 5 Cited by

Patent Information

Application Number
CN202510275230.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-03-10
Publication Date
2025-06-10

AI Technical Summary

Technical Problem

When polishing complex curved parts, traditional CNC grinders cannot perceive and adjust the grinding force in real time, resulting in grinding force fluctuations, resulting in overcut or undercutting, affecting processing quality and production efficiency.

Method used

The surface model of the target polished parts is constructed using high-frequency laser scanner and B-spline surface fitting technology. Combined with a six-dimensional torque sensor and vibration sensor, real-time contact status monitoring is achieved through synchronous trigger pulse sequences, reverse impedance error compensation and joint angle correction are performed, and reverse impedance control of automated polishing equipment is realized.

Benefits of technology

Through high-precision surface modeling and real-time contact status monitoring, precise control of the polishing process is achieved, significantly improving the polishing accuracy and surface finish of complex curved parts, reducing surface defects, and improving product quality and production efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120122558A_ABST
    Figure CN120122558A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of automatic equipment control, in particular to an automatic equipment control system and method and a storage medium. The method comprises the following steps that a high-frequency laser scanner is used for scanning the surface of a target polishing part, B-spline surface fitting is carried out, and a surface fitting surface model is generated; controlling the polishing operation of the polishing industrial robot based on the surface fitting curved surface model, and performing polishing contact state detection to generate contact state point cloud data; based on the contact state point cloud data, grinding expected pose error analysis is carried out, reverse impedance error compensation is carried out, and a reverse impedance compensation amount is obtained; and the joint angle of the polishing industrial robot is corrected through the reverse impedance compensation amount, so that reverse impedance control over the automatic polishing equipment is achieved. By compensating the reverse impedance error of the automatic polishing equipment, accurate control over the automatic polishing process is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of automated equipment control, and particularly to an automated equipment control system, method, and storage medium. Background Art

[0002] Due to their complex geometric shapes and variable curvature characteristics, the polishing of curved surface parts is much more difficult than that of flat or simple curved surfaces. Currently, the polishing of complex curved surface parts mainly relies on numerically controlled grinding machines or industrial robot-assisted polishing systems. Especially for high-precision curved surface parts, such as aeroengine blades, precision optical mirrors, high-end molds, etc., extremely high requirements are placed on polishing accuracy and surface finish. However, the polishing control of traditional numerically controlled grinding machines is mainly based on position control, that is, machining is carried out according to a preset trajectory and feed rate. This control method cannot sense and adjust the contact force in real time during the polishing process, resulting in difficulty in maintaining a constant grinding force when machining complex curved surfaces. Excessive grinding force will cause overcutting on the part surface, forming defects such as tool marks and burns; too small grinding force will cause undercutting, unable to effectively remove the surface material, and failing to achieve the required machining accuracy and surface finish. This problem of improper grinding force control is particularly prominent when dealing with variable curvature curved surfaces, seriously affecting the machining quality and production efficiency of high-precision parts. Summary of the Invention

[0003] Based on this, the present invention provides an automated equipment control system, method, and storage medium to solve at least one of the above technical problems.

[0004] To achieve the above object, an automated equipment control method is applied to an automated polishing equipment. The automated polishing equipment includes a polishing industrial robot, a six-axis force sensor, a vibration sensor, a high-frequency laser scanner, and a control unit. The control unit includes a real-time controller and a driver. The six-axis force sensor, the vibration sensor, and the high-frequency laser scanner are electrically connected to the real-time controller, and the polishing industrial robot is electrically connected to the real-time controller through the driver. The automated equipment control method includes the following steps:

[0005] Step S1: Use the high-frequency laser scanner to scan the surface of the target polishing part to obtain the polished surface point cloud data; perform B-spline surface fitting based on the polished surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control based on the surface fitting surface model to obtain a polishing control instruction;

[0006] Step S2: Control the polishing operation of the polishing industrial robot based on the polishing control instruction, and perform pulse trigger monitoring to generate a synchronous trigger pulse sequence; perform polishing contact state detection based on the synchronous trigger pulse sequence to generate contact state point cloud data;

[0007] Step S3: Analyze the grinding desired pose error based on the contact state point cloud data, and perform reverse impedance error compensation to obtain the reverse impedance compensation amount;

[0008] Step S4: Correct the joint angles of the polishing industrial robot through the reverse impedance compensation amount to achieve the reverse impedance control of the automated polishing equipment.

[0009] Through high-frequency laser scanning and B-spline surface fitting, the present invention accurately constructs the surface model of the target polishing part, which can effectively eliminate the deviations caused by factors such as part manufacturing errors and clamping deformations. The generated polishing control instruction code is not only a simple position instruction, but also includes refined control parameters based on the surface model. This preprocessing method ensures a high degree of consistency between the initial polishing path and the part surface, reducing the risk of excessive grinding in the initial stage. Through the generation and processing of synchronous trigger pulse sequences, the accurate perception of the contact state between the grinding wheel and the part surface during the polishing process is realized. Traditional methods often rely on preset parameters to infer the contact state, while this solution can directly obtain the contact state point cloud data, accurately judge the magnitude, direction, and contact area of the grinding force and other key information, enabling the system to make dynamic adjustments according to the actual grinding situation. Compare the actual contact state with the desired state, and quantify the deviation between the two. Not only consider the force condition of the end effector, but also incorporate the pose deviation caused by this force into the control consideration, which can effectively suppress the over-cutting or under-cutting phenomena caused by excessive or too small grinding force, ensuring the uniform removal of the part surface material. Correct the joint angles of the polishing industrial robot through the reverse impedance compensation amount, realizing the reverse impedance control of the automated polishing equipment, and making fine adjustments within a certain range according to the magnitude and direction of the actual grinding force, thereby maintaining the constancy of the grinding force. Even when facing a variable curvature surface, the robot can adaptively adjust its own posture and motion trajectory to ensure that the grinding wheel always contacts the part surface with an appropriate force. Avoid the problem of grinding force fluctuation caused by the fixed trajectory in position control and inability to cope with curvature changes. Therefore, an automated equipment control method of the present invention realizes the precise control of the automated polishing process through high-precision surface modeling, real-time contact state monitoring, reverse impedance error compensation, and joint angle correction, effectively improving the polishing accuracy and surface finish of complex curved surface parts, reducing surface defects, improving product quality, and at the same time significantly enhancing production efficiency and reducing the cost of manual intervention.

[0010] Preferably, the preliminary polishing control described in step S1 according to the surface fitting surface model includes:

[0011] Evaluate the surface roughness of the machining area according to the surface fitting surface model;

[0012] Obtain the material property data of the polished part; calculate the grinding force on the surface roughness of the machining area based on the initial grinding force range and the material property data of the polished part to generate regional surface grinding force data, where the initial grinding force range is 5N - 10N;

[0013] Extract the surface area curvature according to the surface fitting surface model to generate regional curvature distribution data;

[0014] Set the polishing control points at 2 - 5 points per square millimeter according to the surface roughness of the machining area and the regional curvature distribution data;

[0015] Perform contour grinding path planning for the polishing control points, with the contour interval being 0.2mm - 0.5mm, the fillet radius at the path corner being not less than 0.3mm, the straight-line distance between the path start point and the end point being not less than 10mm, and the total path length error not exceeding ±2.0mm to obtain the grinding path data;

[0016] Perform preliminary polishing control instruction encoding according to the regional surface grinding force data and the grinding path data to obtain the polishing control instruction.

[0017] The present invention evaluates the surface roughness of the machining area according to the surface fitting surface model, making the subsequent grinding force calculation more targeted, avoiding over-grinding of smooth areas and under-grinding of rough areas, improving the uniformity of material removal, and maximizing the retention of the original geometry of the part. Combining with the material property data of the polished part, the solution calculates the grinding force on the surface roughness of the machining area within the initial grinding force range (5N - 10N), avoiding surface damage caused by excessive grinding force, and at the same time ensuring sufficient material removal efficiency. By considering the curvature factor, the grinding force can be more precisely controlled to avoid over-cutting or under-cutting phenomena in areas with large curvature changes. According to the surface roughness of the machining area and the regional curvature distribution data, the polishing control points are set at a density of 2 - 5 points per square millimeter. This adaptive control point setting method ensures sufficient control accuracy in complex curved surface areas, while reducing the control points appropriately in flat areas to improve the polishing efficiency.

[0018] Preferably, the controlling the polishing industrial robot to perform the polishing operation based on the polishing control instruction and performing pulse trigger monitoring in step S2 includes:

[0019] Use the real-time controller to perform control instruction parsing on the polishing control instruction, then control the polishing industrial robot to perform the polishing operation through the driver, and use the vibration sensor to monitor the joint space vibration signal of the polishing industrial robot in real time;

[0020] Extract the vibration signal characteristics of the joint space vibration signal to obtain the joint vibration characteristic data;

[0021] Aggregate joint vibration characteristics based on joint vibration characteristic data, identify the vibration period, and generate the joint vibration period.

[0022] Based on a preset polishing scan threshold, count the number of trigger times for vibration period scanning through the joint vibration period to obtain the number of triggers per period.

[0023] Calculate the relative time delay of each trigger point according to the number of triggers per period and the joint vibration characteristic data to obtain the trigger time delay sequence.

[0024] Evaluate the polishing steady state based on the joint vibration characteristic data to obtain the polishing stability index.

[0025] Obtain real-time clock data; perform adaptive monitoring trigger pulse processing on the trigger time delay sequence based on the polishing stability index and the real-time clock data to generate a synchronous trigger pulse sequence.

[0026] The present invention uses a vibration sensor to monitor the joint space vibration signal of a polishing industrial robot in real time. By monitoring the vibration of the joint, the change of the contact state between the grinding wheel and the part surface can be indirectly reflected, such as the change of the contact force, the change of the friction coefficient, etc. By extracting key vibration characteristics, such as amplitude, frequency, energy, etc., the state change during the polishing process can be more clearly reflected. Aggregating joint vibration characteristics based on joint vibration characteristic data, the identification of the vibration period can help the system judge the stability of the polishing process. Based on a preset polishing scan threshold, counting the number of trigger times for vibration period scanning through the joint vibration period can help the system judge the density of the vibration signal, so as to infer the severity of the contact state during the polishing process. Calculating the relative time delay of each trigger point according to the number of triggers per period and the joint vibration characteristic data can more accurately determine the optimal moment of the trigger pulse, so as to obtain more representative contact state information. Evaluating the polishing steady state based on the joint vibration characteristic data enables the system to quantitatively evaluate the stability of the polishing process. This helps the system judge whether the polishing process is in the best state and make timely adjustments to avoid the decline of product quality caused by unstable polishing. Performing adaptive monitoring trigger pulse processing on the trigger time delay sequence based on the polishing stability index and the real-time clock data can dynamically adjust the frequency and timing of the trigger pulse according to the actual state of the polishing process, so as to ensure high-quality contact state data can be obtained under different working conditions.

[0027] Preferably, the polishing contact state detection according to the synchronous trigger pulse sequence in step S2 includes:

[0028] Based on the synchronous trigger pulse sequence, control a high-frequency laser scanner to perform trigger-type polishing surface scanning to generate a high-frequency scanned original point cloud.

[0029] Control the six - dimensional torque sensor to measure the polishing force torque of the end effector of the polishing industrial robot according to the synchronous trigger pulse sequence, and generate polishing force torque data;

[0030] Filter and calibrate the data according to the high - frequency scanned original point cloud, and perform point cloud segmentation to obtain single - frame point cloud data;

[0031] Synchronize the time stamps of the polishing force torque data through the single - frame point cloud data to generate synchronized force - point cloud data;

[0032] Identify the polishing contact state of the synchronized force - point cloud data to generate contact state point cloud data.

[0033] Based on the synchronous trigger pulse sequence, the present invention simultaneously controls the high - frequency laser scanner to perform trigger - type polishing surface scanning and controls the six - dimensional torque sensor to measure the polishing force torque. The high - frequency laser scanner can provide the geometric information of the part surface, while the six - dimensional torque sensor can provide the torque information generated during the grinding process. Filter and calibrate the high - frequency scanned original point cloud. Data filtering can remove noise and outliers, improving the accuracy of the point cloud data; data calibration can eliminate the deviation caused by factors such as sensor error and assembly error, ensuring the accuracy of the point cloud data. The point cloud data after filtering and calibration can more truly reflect the geometric shape of the part surface. Since the high - frequency laser scanner and the six - dimensional torque sensor work independently, there is a time deviation in their data. Through time - stamp synchronization, this deviation can be eliminated, ensuring the consistency of force and point cloud data in time. This enables the system to more accurately analyze the relationship between torque and the geometric shape of the part surface, and thus more precisely judge the contact state. By comprehensively analyzing the force and point cloud data, the system can judge information such as whether the grinding wheel is in contact with the part surface, the size of the contact area, and the direction of the contact force.

[0034] Preferably, the identifying the polishing contact state of the synchronized force - point cloud data includes:

[0035] Extract the origin of the sensor Cartesian coordinate system according to the polishing force torque data, and calculate the pose of the polishing head to generate polishing head pose data;

[0036] Calculate the distance between the polishing head and the workpiece according to the polishing head pose data to obtain distance change data;

[0037] Calculate the torque according to the polishing force torque data, and respectively obtain the force component value in the Z - axis direction and the torque component values around the X and Y axes;

[0038] Calculate the force change rate according to the force component value in the Z - axis direction to generate the Z - axis force change rate;

[0039] When the distance change data is less than 0.05 mm, the Z-axis direction force component value is greater than 0.5 N, the square root of the sum of the squares of the moment component values about the X and Y axes is greater than 0.05 Nm, and the Z-axis force change rate is less than 0.2 N / s, the synchronous force-point cloud data is marked as contact point cloud data;

[0040] When the distance change data is greater than 0.1 mm, the Z-axis direction force component value is less than 0.1 N, and the square root of the sum of the squares of the moment component values about the X and Y axes is less than 0.01 Nm, the synchronous force-point cloud data is marked as off-contact point cloud data; otherwise, the synchronous force-point cloud data is marked as critical contact point cloud data;

[0041] Integrate the contact point cloud data, off-contact point cloud data, and critical contact point cloud data in terms of point cloud state to obtain contact state point cloud data.

[0042] The present invention extracts the origin of the sensor Cartesian coordinate system based on the polishing acting moment data, calculates the pose of the polishing head, and understands the relative position relationship between the grinding wheel and the part. Calculate the distance between the polishing head and the workpiece according to the polishing head pose data, which reflects the proximity between the grinding wheel and the part. Calculate the moment according to the polishing acting moment data to obtain the Z-axis direction force component value and the moment component values about the X and Y axes respectively. The Z-axis direction force component value can reflect the magnitude of the normal pressure between the grinding wheel and the part, while the moment component values about the X and Y axes can reflect the magnitude of the frictional force between the grinding wheel and the part. Calculate the force change rate according to the Z-axis direction force component value, which reflects the change speed of the contact force between the grinding wheel and the part, and helps to judge the stability of the contact state. Identify the true contact state, integrate the contact point cloud data, off-contact point cloud data, and critical contact point cloud data in terms of point cloud state, and comprehensively understand the change of the contact state during the polishing process.

[0043] Preferably, step S3 includes the following steps:

[0044] Step S31: Extract the contact state at the current moment according to the contact state point cloud data;

[0045] Step S32: Match the impedance parameters for the contact state at the current moment through a preset impedance adjustment rule to obtain real-time impedance adjustment parameters;

[0046] Step S33: Analyze the surface gradient at the point at the current moment through the surface fitting surface model, and then perform processing on the surface normal vector to obtain the real-time surface normal vector;

[0047] Step S34: Use the real-time surface normal vector to perform current normal pose analysis on the contact state point cloud data to generate current normal pose data;

[0048] Step S35: Analyze the grinding desired pose error of the surface fitting surface model based on the current normal pose data to obtain the grinding normal pose error;

[0049] Step S36: Perform normal force / tangential force impedance error compensation according to the grinding normal pose error and the real-time impedance adjustment parameters, and perform inverse Jacobian matrix solution processing to obtain the reverse impedance compensation amount.

[0050] The present invention extracts the current contact state from the contact state point cloud data. Only by accurately understanding the current contact state can a suitable control strategy be formulated. By matching the impedance parameters for the current contact state according to the preset impedance adjustment rules, the impedance parameters, such as stiffness and damping, can be adaptively adjusted according to different contact states, so as to achieve a more compliant polishing effect. For example, in the case of relatively intense contact, the stiffness can be reduced and the damping can be increased to reduce the impact; while in the case of relatively stable contact, the stiffness can be increased to improve the accuracy. Using the real-time surface normal vector to perform the current normal pose analysis on the contact state point cloud data can more comprehensively evaluate the grinding state. Analyzing the grinding desired pose error of the surface fitting surface model based on the current normal pose data can accurately reflect the deviation between the actual grinding state and the desired state.

[0051] Preferably, step S34 includes the following steps:

[0052] Step S341: Extract the three-dimensional coordinates of the current polishing point according to the contact state point cloud data;

[0053] Step S342: Use the real-time surface normal vector as the direction vector, and take the three-dimensional coordinates of the current polishing point as the origin to construct the normal axis at the current moment;

[0054] Step S343: Calculate the perpendicular projection point of the three-dimensional coordinates of the current polishing point through the normal axis at the current moment to obtain the current normal position;

[0055] Step S344: Extract the three-dimensional coordinates of the previous polishing point according to the contact state point cloud data;

[0056] Step S345: Calculate the perpendicular projection point of the three-dimensional coordinates of the previous polishing point through the normal axis at the current moment to obtain the normal position at the previous moment;

[0057] Step S346: Calculate the normal velocity based on the current normal position and the normal position at the previous moment to obtain the current normal velocity;

[0058] Step S347: Combine the state variables according to the current normal position and the current normal velocity to generate the current normal pose data.

[0059] The present invention extracts the three-dimensional coordinates of the current polishing point from the contact state point cloud data, which represents the actual contact position between the grinding wheel and the part surface. Taking the real-time surface normal vector as the direction vector and the three-dimensional coordinates of the current polishing point as the origin, the normal axis at the current moment is constructed. The construction of the normal axis decomposes the motion in the three-dimensional space into the motion along the normal direction and the motion perpendicular to the normal direction. By calculating the perpendicular projection point of the three-dimensional coordinates of the current polishing point through the normal axis at the current moment, the approaching degree between the grinding wheel and the part surface in the normal direction is reflected. The three-dimensional coordinates of the polishing point at the previous moment are extracted from the contact state point cloud data, and the perpendicular projection point of the three-dimensional coordinates of the polishing point at the previous moment is calculated through the normal axis at the current moment to obtain the normal position at the previous moment. With the normal position information at the current moment and the previous moment, the scheme can calculate the normal velocity. The normal velocity reflects the approaching or separating velocity between the grinding wheel and the part surface in the normal direction, and the system can better understand the dynamic characteristics during the polishing process. By combining the current normal position and the current normal velocity, the motion state of the polishing point in the normal direction can be more comprehensively described.

[0060] Preferably, step S4 includes the following steps:

[0061] Step S41: Obtain the current robot joint angle value;

[0062] Step S42: Perform joint space compensation amount conversion on the current robot joint angle value through the reverse impedance compensation amount to obtain the joint space compensation amount;

[0063] Step S43: Correct the spatial joint angle according to the joint space compensation amount and synthesize the joint control command to obtain the joint drive control signal;

[0064] Step S44: Send the joint drive control signal to the driver through the communication bus to control the correction of each joint angle of the polishing industrial robot, so as to realize the reverse impedance control of the automatic polishing equipment.

[0065] The present invention converts the current robot joint angle value through the reverse impedance compensation amount to convert the impedance compensation amount in the Cartesian space into the compensation amount in the robot joint space, so that the robot can perform actual motion according to the compensation amount. Correcting the spatial joint angle according to the joint space compensation amount can convert the compensation amount into control instructions that the robot can recognize and execute. The joint control instructions need to consider the dynamic characteristics and motion constraints of the robot to ensure the smoothness and safety of the motion. Sending the joint drive control signal to the driver through the communication bus can actually apply the calculated control instructions to the robot, thereby realizing precise control of the polishing process. By adjusting the joint angle of the robot in real time, the polishing wheel can always contact the workpiece surface with appropriate force and posture, ensuring the constancy of the grinding force and the surface finish, and at the same time avoiding defects such as over-cutting or under-cutting.

[0066] Preferably, the present invention also provides an automated equipment control system that executes the automated equipment control method described above. The automated equipment control system includes:

[0067] A polishing instruction generation module for scanning the surface of the target polishing part using a high-frequency laser scanner to obtain polished surface point cloud data; performing B-spline surface fitting according to the polished surface point cloud data to generate a surface fitting surface model; performing preliminary polishing control according to the surface fitting surface model to obtain polishing control instructions;

[0068] A polishing state monitoring module for controlling the polishing operation of the polishing industrial robot based on the polishing control instructions, and performing pulse trigger monitoring to generate a synchronous trigger pulse sequence; detecting the polishing contact state according to the synchronous trigger pulse sequence to generate contact state point cloud data;

[0069] An impedance error analysis module for analyzing the grinding desired pose error based on the contact state point cloud data and performing impedance error compensation to obtain a reverse impedance compensation amount;

[0070] An equipment joint impedance control module for correcting the joint angle of the polishing industrial robot through the reverse impedance compensation amount to achieve reverse impedance control of the automated polishing equipment.

[0071] Preferably, the present invention also provides a computer-readable storage medium storing a computer program, and when the computer program is executed, it implements the automated equipment control method described in any one of the above. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a schematic flow chart of the steps of an automated equipment control method of the present invention;

[0073] Figure 2 isFigure 1 A schematic diagram of a detailed implementation process of performing preliminary polishing control according to a surface fitting surface model in step S1;

[0074] Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0075] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0076] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0077] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0078] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0079] To achieve this, please refer to Figures 1 to 3 The present invention provides an automated equipment control method, which is applied to automated polishing equipment. The automated polishing equipment includes a polishing industrial robot, a six-dimensional torque sensor, a vibration sensor, a high-frequency laser scanner, and a control unit. The control unit includes a real-time controller and a driver. The six-dimensional torque sensor, the vibration sensor, and the high-frequency laser scanner are electrically connected to the real-time controller. The polishing industrial robot is electrically connected to the real-time controller through the driver. The automated equipment control method includes the following steps:

[0080] Step S1: Scan the surface of the target polishing part with a high-frequency laser scanner to obtain polishing surface point cloud data; perform B-spline surface fitting according to the polishing surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control according to the surface fitting surface model to obtain a polishing control instruction;

[0081] Step S2: Based on the polishing control instruction, the polishing industrial robot is controlled to perform polishing operation, and pulse trigger monitoring is performed to generate a synchronous trigger pulse sequence; polishing contact state detection is performed according to the synchronous trigger pulse sequence to generate contact state point cloud data;

[0082] Step S3: performing grinding desired posture error analysis based on the contact state point cloud data, and performing reverse impedance error compensation to obtain a reverse impedance compensation amount;

[0083] Step S4: Correcting the joint angle of the polishing industrial robot by using the reverse impedance compensation amount to achieve reverse impedance control of the automated polishing equipment.

[0084] In an embodiment of the present invention, the automation equipment control method comprises the following steps:

[0085] Step S1: Scan the surface of the target polishing part with a high-frequency laser scanner to obtain polishing surface point cloud data; perform B-spline surface fitting according to the polishing surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control according to the surface fitting surface model to obtain a polishing control instruction;

[0086] In an embodiment of the present invention, a structured light laser scanner (e.g., Keyence LJ-V7000, scanning frequency 64kHz) is used to scan a turbine blade made of aluminum alloy 6061-T6. The blade is fixed on a fixture, and the position and angle of the laser scanner are adjusted so that it can completely scan the surface to be polished of the blade. The scanner is started and scanned in line scanning mode to obtain a series of contour line data. Each contour line contains multiple points, and each point contains two-dimensional coordinate information (x, y) and laser intensity information. All the contour line data obtained by scanning are combined together in the scanning order to form raw point cloud data. The raw point cloud data is preprocessed, including removing outliers (using a statistical filter, setting the neighborhood point count threshold to 50, and the standard deviation multiple threshold to 1) and filtering and smoothing (using a moving average filter, with a window size of 5x5). The B-spline surface fitting method based on the least squares method is used to reconstruct the surface of the preprocessed point cloud data. The control point grid is set to 30x30, and the order of the B-spline surface is set to 3rd order. An adaptive node vector generation algorithm is used to adjust the node vector according to the distribution density of point cloud data. The coordinates of the control points of the B-spline surface are obtained by solving the least squares problem, thereby generating a parametric surface model that can accurately represent the geometric shape of the blade surface. Based on the surface model and the preprocessed point cloud data, polishing control instructions are encoded. First, the surface roughness of the processing area is evaluated: the normal vector change rate of each data point in its local neighborhood (k = 15 nearest neighbors) is calculated as the roughness evaluation value. Secondly, the material property database is queried to obtain the Young's modulus (68.9GPa), Poisson's ratio (0.33) and hardness (95HB) of aluminum alloy 6061-T6. Then, based on the initial grinding force range (5N-10N) and Hertz contact theory, the expected grinding force corresponding to each roughness evaluation value is calculated. Next, the principal curvature at each parameter point on the surface model is calculated, and the product of the two principal curvatures is taken as the curvature value of the point. According to the roughness and curvature distribution, the polishing control points are set at a density of 3 points per square millimeter. The contour offset algorithm is used to generate the grinding path, with a contour interval of 0.4 mm, a corner radius of 0.5 mm, and a straight-line distance between the path start and end greater than 15 mm. Finally, the expected grinding force at each control point and the grinding path data are combined to generate preliminary polishing control instructions.

[0087] Step S2: Based on the polishing control instruction, the polishing industrial robot is controlled to perform polishing operation, and pulse trigger monitoring is performed to generate a synchronous trigger pulse sequence; polishing contact state detection is performed according to the synchronous trigger pulse sequence to generate contact state point cloud data;

[0088] In the embodiment of the present invention, a real-time controller based on EtherCAT bus (cycle period 1ms) is used to control an ABB IRB 1200 industrial robot to perform polishing operations. A rubber polishing head and an ATIGamma six-dimensional torque sensor are installed at the end of the robot. The real-time controller receives the polishing control instruction generated in step S1, and converts the position and posture instructions into the joint angle instructions of the robot, and passes the expected normal force instruction to the impedance controller (the parameters are set as: M=1kg, B=100Ns / m, K=1000N / m). Piezoelectric acceleration sensors (PCB 352C33, sampling frequency 10kHz) are installed on each joint of the robot for real-time monitoring of joint vibration signals. The controller sends the joint angle instruction to the robot driver via the EtherCAT bus to drive the robot to move along a predetermined path. At the same time, the impedance controller calculates the corrected position instruction based on the actual normal force and the expected normal force fed back by the torque sensor, and sends it to the driver to achieve precise control of the polishing force. The vibration signal collected by the acceleration sensor was subjected to short-time Fourier transform (Hanning window, window length 1024 points, overlap rate 50%), and the RMS value, main frequency, spectrum energy and spectrum entropy were extracted as joint vibration feature data. The PCA method was used to reduce the dimension of the feature data, and the autocorrelation function method was used to identify the vibration cycle. The polishing scan threshold was set to 3 times the standard deviation of the feature data within the vibration cycle, and the number of times the feature data exceeded the threshold in each vibration cycle was counted. The relative time delay of each trigger point relative to the starting point of the vibration cycle was calculated to obtain the trigger delay sequence. The RMS value of the feature data in each vibration cycle was calculated as the polishing stability index. According to the stability index and real-time clock data, the trigger delay sequence was adaptively processed to generate a synchronous trigger pulse sequence. The synchronous trigger pulse sequence was sent to the high-frequency laser scanner (Keyence LJ-V7000) and the data acquisition card of the torque sensor at the same time. When the laser scanner received each trigger pulse, it scanned the contact area once to generate a high-frequency scan raw point cloud. When the torque sensor received each trigger pulse, it collected torque data once. The scanned point cloud is subjected to median filtering (5x5 window) and coordinate transformation, and is segmented into single-frame point clouds according to the trigger pulse time interval. The single-frame point cloud data and the corresponding torque data are added with the same timestamp to form synchronized force-point cloud data. According to the normal force threshold (1N), torque threshold (0.1Nm) and point cloud distance threshold (0.1mm), the synchronized force-point cloud data is marked as "contact", "disconnection" or "critical contact" to generate contact state point cloud data.

[0089] Step S3: performing grinding desired posture error analysis based on the contact state point cloud data, and performing reverse impedance error compensation to obtain a reverse impedance compensation amount;

[0090] In an embodiment of the present invention, the contact state ("contact", "disconnection" or "critical contact") at the current moment is extracted. According to the preset impedance adjustment rule library, the corresponding impedance parameters (K, B) are selected: contact state (K = 1500N / m, B = 120Ns / m); disconnection state (K = 500N / m, B = 50Ns / m); critical contact state (K = 1000N / m, B = 150Ns / m). Based on the surface fitting surface model generated in step S1, the surface normal vector n at the polishing point at the current moment is calculated. The specific method is: find the point closest to the current polishing point on the surface model, calculate the tangent vectors of the two main directions at the point, and then cross-multiply to obtain the normal vector and normalize it. Extract the three-dimensional coordinates of the current polishing point and the polishing point at the previous moment from the contact state point cloud data. Take the current polishing point as the origin and the real-time surface normal vector as the direction vector to construct the normal axis at the current moment. Project the current polishing point and the polishing point at the previous moment vertically onto the normal axis to obtain the current normal position and the normal position at the previous moment. Calculate the distance difference between the two projected points and divide it by the time interval to obtain the current normal velocity. Combine the current normal position (3D coordinate) and the current normal velocity (scalar) to form the current normal pose data. Find the nearest point to the current polishing point on the surface model and calculate the normal vector at that point. Calculate the angle between the expected normal vector in the current normal pose data and the normal vector as the normal angle error. Calculate the distance between the current polishing point and the nearest point as the normal position error. Convert the normal angle error to the expected torque compensation (angle stiffness coefficient 10Nm / rad), and convert the normal position error to the expected force compensation (stiffness K comes from the impedance adjustment rule). Calculate the impedance torque of the robot end in Cartesian space: F_i=[0,0,ΔF,ΔMx,ΔMy,0]. Using the Jacobian matrix J of the robot, F_i is converted into the compensation amount Δq=J^T×F_i in the joint space (solved using pseudo-inverse or damped least squares method). Δq is the reverse impedance compensation amount.

[0091] Step S4: Correcting the joint angle of the polishing industrial robot by using the reverse impedance compensation amount to achieve reverse impedance control of the automated polishing equipment.

[0092] In an embodiment of the present invention, the real-time controller obtains the current angle value q_current of each joint of the robot through the EtherCAT bus (measured using a 16-bit absolute encoder, unit: radian). Receive the reverse impedance compensation amount Δq calculated in step S3. Since Δq is already the compensation amount of the joint space, it is directly added to q_current to obtain the corrected joint angle q_corrected = q_current + Δq. Limit each element of q_corrected within the range of motion of the corresponding joint of the robot to prevent exceeding the joint limit. Convert the corrected joint angle q_corrected into an encoder count value: counts = (q_corrected / (2π)) × encoder_resolution, where encoder_resolution is the resolution of the encoder. Combine the encoder count values ​​of each joint into a complete joint control instruction data packet (following the EtherCAT protocol format). The real-time controller sends the joint drive control signal to each joint driver of the robot through the EtherCAT bus (master-slave communication, cycle period 1ms). After the driver receives the control instruction, it performs a CRC check. After the verification is passed, the PID controller inside the driver calculates the motor control voltage based on the received encoder count value (expected joint angle) and the actual encoder feedback value (actual joint angle), and drives the joint motor to rotate so that the joint angle approaches the expected value. Through this closed-loop control, the robot joint angle can be accurately corrected, thereby realizing the reverse impedance control of the automated polishing equipment, ensuring that the polishing head maintains a stable contact force and correct posture with the workpiece surface.

[0093] As an example of the present invention, refer to Figure 2 As shown, Figure 1 The schematic diagram of the implementation process of performing preliminary polishing control according to the surface fitting curved surface model in step S1, in this example, performing preliminary polishing control according to the surface fitting curved surface model includes:

[0094] S11: Evaluate the surface roughness of the machining area based on the surface fitting surface model;

[0095] In an embodiment of the present invention, a B-spline surface fitting method based on the least squares method is used to reconstruct the surface of the preprocessed polished surface point cloud data. During the reconstruction process, the control point grid is set to 20x20, and the order of the B-spline surface is set to 3. After the fitting is completed, a parameterized surface model that can accurately represent the geometric shape of the polished surface is obtained. Using this surface model, the normal vector change rate of each data point in its local neighborhood is calculated. The specific operation is: for each point Pi in the point cloud, select its k-nearest neighbor (k=10) point set {P1, P2,..., P10}. Calculate the normal vectors of point Pi and its k-nearest neighbor points on the fitting surface respectively, and obtain the normal vector set {N0, N1, N2,..., N10}. Calculate the angles between N0 and N1 to N10 respectively, and obtain 10 angle values. Calculate the standard deviation of these 10 angle values ​​as the roughness evaluation value Ri at point Pi. Traverse all point cloud data points to obtain the surface roughness distribution data {R1, R2, ..., Rn} of the entire processing area.

[0096] S12: Obtaining polishing part material property data; performing grinding force calculation on the surface roughness of the processing area based on the initial grinding force range and the polishing part material property data to generate regional surface grinding force data, wherein the initial grinding force range is 5N-10N;

[0097] In an embodiment of the present invention, a material property database is established, which is obtained through experimental testing and stores property parameters of different materials, including mechanical performance parameters such as Young's modulus, Poisson's ratio, yield strength, and hardness. For example, for the polishing object this time, aluminum alloy 6061-T6, by querying the database, it is obtained that its Young's modulus is 68.9GPa, Poisson's ratio is 0.33, yield strength is 276MPa, and hardness (Brinell hardness) is 95HB. For each roughness evaluation value Ri, the required grinding force is calculated according to the Hertz contact theory. The formula of the Hertz contact theory is: F = (4 / 3) × E × R^(1 / 2) × δ^(3 / 2), where F is the grinding force, E is the equivalent Young's modulus, R is the equivalent radius of curvature of the contact area, and δ is the normal indentation depth. First, the equivalent Young's modulus E is calculated based on the Young's modulus and Poisson's ratio of aluminum alloy 6061-T6 and the Young's modulus and Poisson's ratio of the polishing tool (assuming it is made of rubber and the Young's modulus and Poisson's ratio are known). Secondly, the equivalent radius of curvature R of the contact area is estimated based on the roughness evaluation value Ri. Here, the empirical formula R=k1×Ri is used, where k1 is the empirical coefficient, which is 0.1 here. Finally, according to the initial grinding force range of 5N-10N, the search range of the initial normal indentation depth δ is set. Through iterative calculation, the δ value that satisfies the Hertz contact theory formula is found, and the corresponding grinding force F is calculated. The F value is limited to the range of 5N-10N, and the boundary value is taken if it exceeds the range. Traverse all roughness evaluation values ​​to obtain the surface grinding force data {F1, F2,..., Fn} of the entire processing area.

[0098] S13: extracting surface regional curvature according to the surface fitting surface model to generate regional curvature distribution data;

[0099] In an embodiment of the present invention, the first basic form coefficients E, F, G and the second basic form coefficients L, M, N of the surface at each parameter point (u, v) are calculated. Then, the Gauss-Weingarten equation is used to calculate the principal curvatures k1 and k2 of the surface at that point, and the formula is: k1, k2 = H ± √ (H^2-K), where H is the mean curvature, K is the Gaussian curvature, and H and K can be calculated from E, F, G, L, M, N. After calculating the principal curvature at each parameter point, the product of the two principal curvatures is defined as the curvature value of the point. Traverse all parameter points to obtain the curvature distribution data {C1, C2, ..., Cm} of the entire processing area.

[0100] S14: setting polishing control points at 2-5 points per square millimeter according to the surface roughness of the processing area and the regional curvature distribution data;

[0101] In the embodiments of the present invention, the machining area is divided into grids of 1 mm x 1 mm. For each grid cell, the average roughness value Ra and the average curvature value Ca of all the data points inside it are calculated. According to the values of Ra and Ca, the number of control points within the grid cell is determined. The specific rules are as follows: If Ra is greater than the preset threshold R_threshold (for example, R_threshold = 0.5 μm) and Ca is greater than the preset threshold C_threshold (for example, C_threshold = 0.1), then 5 control points are set within the grid cell; if Ra is greater than R_threshold or Ca is greater than C_threshold, then 4 control points are set; if both Ra and Ca are less than their respective thresholds, then 2 control points are set; in other cases, 3 control points are set. The control points are evenly distributed within the grid cell.

[0102] S15: Perform contour grinding path planning for the polishing control points, with the contour interval being 0.2 mm - 0.5 mm, the fillet radius at the path turning corners being not less than 0.3 mm, the straight-line distance between the path start point and the end point being not less than 10 mm, and the total path length error not exceeding ±2.0 mm, to obtain the grinding path data;

[0103] In the embodiments of the present invention, according to the determined control points, an irregular triangular network is constructed. Then, along the sides of the triangular network, a series of contour lines are generated at the contour interval (for example, 0.3 mm). When generating the contour lines, the sharp corners need to be filleted, and the fillet radius is set to 0.4 mm. To ensure the machining efficiency, the straight-line distance between the start point and the end point of each contour line should be greater than 10 mm. After generating a complete contour line, calculate its total length and compare it with the target length. If the error exceeds ±2.0 mm, then adjust the contour line, such as adding or reducing some line segments, until the length error requirement is met. Finally, a series of contour paths that meet the requirements are obtained, which are the grinding path data.

[0104] S16: Perform preliminary polishing control instruction encoding based on the regional surface grinding force data and the grinding path data to obtain the polishing control instructions.

[0105] In an embodiment of the present invention, the grinding path data (a series of spatial coordinate points) is associated with the corresponding grinding force data (the desired grinding force at each coordinate point). Using the method of linear interpolation, the desired grinding force at each coordinate point on the grinding path is calculated. For example, for a point P on the path, find the two nearest data points P1 and P2, and the corresponding grinding forces F1 and F2. According to the distance relationship between P, P1, and P2, the desired grinding force Fp at point P is calculated by linear interpolation. Arrange each path point and its corresponding desired grinding force Fp in chronological order to generate a preliminary polishing control instruction. This instruction includes position information (x, y, z coordinates) and force information (Fx, Fy, Fz, where Fx and Fy are tangential forces, and Fz is the normal force, i.e., Fp). This control instruction will be sent to the real-time controller of the automated equipment.

[0106] Preferably, in step S2, controlling the polishing operation of the polishing industrial robot based on the polishing control instruction and performing pulse trigger monitoring includes:

[0107] Using the real-time controller to parse the polishing control instruction, and then controlling the polishing industrial robot to perform the polishing operation through the driver, and using the vibration sensor to monitor the joint space vibration signal of the polishing industrial robot in real time;

[0108] Extracting the vibration signal characteristics of the joint space vibration signal to obtain joint vibration characteristic data;

[0109] Aggregating the joint vibration characteristics according to the joint vibration characteristic data, and performing vibration period identification to generate a joint vibration period;

[0110] Based on a preset polishing scan threshold, counting the number of vibration period scan triggers through the joint vibration period to obtain the number of triggers per period;

[0111] Calculating the relative time delay of each trigger point according to the number of triggers per period and the joint vibration characteristic data to obtain a trigger time delay sequence;

[0112] Evaluating the polishing stable state according to the joint vibration characteristic data to obtain a polishing stability index;

[0113] Obtaining real-time clock data; performing adaptive monitoring trigger pulse processing on the trigger time delay sequence based on the polishing stability index and the real-time clock data to generate a synchronous trigger pulse sequence.

[0114] In the embodiments of the present invention, a real-time controller based on the EtherCAT bus is adopted. The main frequency of the controller is set to 1 GHz, and the cycle period is set to 1 ms. The controller receives polishing control instructions (including position information and force information). The instruction parsing module inside the controller converts the position information (x, y, z coordinates) into the desired angles of each joint of the industrial robot, and converts the force information (Fx, Fy, Fz) into the desired normal force of the end effector. These expected values are sent to the joint drivers of the industrial robot (e.g., ABB IRB 1200) through the EtherCAT bus. The driver adopts the PID control algorithm to control each joint motor to move according to the desired angle. At the same time, a six-axis force sensor (e.g., ATI Gamma) installed on the end effector measures the actual force on the end effector in real time and feeds it back to the real-time controller. The real-time controller sends the expected normal force and the actual normal force to the driver to achieve precise control of the polishing force. At the same time, a piezoelectric acceleration sensor (e.g., PCB 352C33) is installed at each joint of the industrial robot to monitor the vibration acceleration signal of the joint in real time. The sampling frequency of the sensor is set to 10 kHz. The short-time Fourier transform (STFT) is used to perform time-frequency analysis on the preprocessed signal. The window function of the STFT is selected as the Hanning window, the window length is set to 1024 sampling points, and the overlap rate is set to 50%. The Fourier transform is performed on the signal within each time window to obtain its spectrum. The following features are extracted from the spectrum: (1) the root mean square value (RMS) of the vibration signal; (2) the main frequency of the vibration signal; (3) the spectral energy of the vibration signal; (4) the spectral entropy of the vibration signal. These features are combined into a feature vector as the joint vibration feature data of the joint within this time window. The principal component analysis (PCA) method is used to reduce the dimension of the aggregated feature data. The first three principal components are retained, and these principal components explain more than 95% of the total variance. The dimension-reduced feature data is used as the new joint vibration feature data. The autocorrelation function method is used to identify the vibration period, and the delay time corresponding to the first peak is found, which is the joint vibration period T. The preset polishing scan threshold is set to 3 times the standard deviation of the joint vibration feature data within its vibration period T. Within each vibration period T, the joint vibration feature data is scanned. If the value of the feature data exceeds the polishing scan threshold, the counter is incremented by 1. The entire vibration period T is traversed to obtain the number of single-cycle triggers N. The time ti when the trigger point occurs and the start time t0 of this vibration period are recorded. The relative time delay Δti = ti - t0 is calculated. The relative time delays of all trigger points are arranged in chronological order to obtain the trigger time delay sequence {Δt1, Δt2,..., ΔtN}. A stability evaluation method based on vibration energy is adopted. The root mean square value (RMS) of the joint vibration feature data within each vibration period T is calculated. The RMS value is used as the polishing stability index. The smaller the RMS value, the more stable the polishing process.At the starting moment t0 of each vibration cycle, add the corresponding relative time delay Δti to obtain the trigger moment Ti = t0 + Δti. At the moment Ti, a pulse signal is generated, and the pulse width is set to 1 millisecond. If the RMS value is greater than the threshold, it is considered that the polishing process is unstable. According to the polishing stability index, adjust the frequency and amplitude of the trigger pulse. If the polishing stability index is high, it indicates that the polishing process is stable, and the frequency of the trigger pulse can be appropriately reduced to reduce the interference to the system; if the polishing stability index is low, it indicates that the polishing process is unstable, and the frequency of the trigger pulse needs to be increased to respond to changes more quickly. For example, a linear relationship between the trigger pulse frequency and the polishing stability index can be established, that is, trigger pulse frequency = k × (1 - polishing stability index), where k is a proportionality coefficient that needs to be adjusted according to the actual situation. Next, according to the trigger delay sequence, determine the exact time points of each trigger pulse to form a synchronous trigger pulse sequence.

[0115] Preferably, the polishing contact state detection according to the synchronous trigger pulse sequence in step S2 includes:

[0116] Based on the synchronous trigger pulse sequence, control the high-frequency laser scanner to perform trigger-type polishing surface scanning to generate high-frequency scanned original point cloud;

[0117] According to the synchronous trigger pulse sequence, control the six-dimensional torque sensor to measure the polishing acting torque of the end effector of the polishing industrial robot to generate polishing acting torque data;

[0118] Perform data filtering and calibration on the high-frequency scanned original point cloud, and perform point cloud segmentation to obtain single-frame point cloud data;

[0119] Synchronize the time stamps of the polishing acting torque data through the single-frame point cloud data to generate synchronous force-point cloud data;

[0120] Identify the polishing contact state of the synchronous force-point cloud data to generate contact state point cloud data.

[0121] In the embodiments of the present invention, a structured light laser scanner (e.g., Keyence LJ-V7000) is used, and its scanning frequency is set to 64 kHz. A synchronous trigger pulse sequence is input into the external trigger interface of the laser scanner. When the laser scanner receives each trigger pulse, it immediately triggers a scan to obtain the contour line data of the contact area between the polishing tool and the workpiece. Since the trigger pulse is synchronized with the joint vibration of the industrial robot, the laser scanner can capture the contact state at a specific moment within each vibration cycle. The contour line data obtained from each scan contains multiple points, and each point contains two-dimensional coordinate information (x, y) and laser intensity information. All the obtained contour line data are combined together in chronological order to form the original high-frequency scanned point cloud data. A six-axis force sensor (e.g., ATI Gamma) is used, which is installed on the end effector of the industrial robot. A synchronous trigger pulse sequence is input into the external trigger interface of the data acquisition card of the force sensor. When the force sensor receives each trigger pulse, it immediately triggers a data acquisition to obtain the polishing force (Fx, Fy, Fz) and torque (Mx, My, Mz) data at the current moment. Since the trigger pulse is synchronized with the joint vibration of the industrial robot, the force sensor can capture the torque information at a specific moment within each vibration cycle. All the acquired torque data are combined together in chronological order to form the polishing force and torque data. Preprocessing is performed on the original high-frequency scanned point cloud data. First, the median filtering algorithm is used to remove the isolated noise points in the point cloud, and the filter window size is set to 5x5. Then, according to the calibration parameters of the laser scanner, distortion correction and coordinate system conversion are performed on the point cloud data, and the point cloud data is converted to the base coordinate system of the industrial robot. Since each scan obtains a contour line, the continuous contour line data needs to be segmented into single-frame point cloud data. According to the time interval of the synchronous trigger pulse, the original point cloud data is segmented into multiple segments, and each segment corresponds to a trigger pulse, that is, one frame of point cloud data. Since both the laser scanner and the force sensor are triggered by the same synchronous trigger pulse sequence, the data they acquire has the same time reference. The same timestamp is added to the single-frame point cloud data and the corresponding torque data. The accuracy of the timestamp is set to 1 millisecond. The single-frame point cloud data and the torque data with the same timestamp are combined together to form the synchronous force-point cloud data. A contact state recognition method based on torque threshold and point cloud distance is adopted. First, a normal force threshold Fz_threshold (e.g., Fz_threshold = 1 N) and a torque threshold M_threshold (e.g., M_threshold = 0.1 Nm) are set. If the normal force Fz in the synchronous force-point cloud data is greater than Fz_threshold, and the absolute value of at least one torque component (Mx, My, Mz) is greater than M_threshold, it is preliminarily determined that the contact state exists.Then, calculate the distance from each point in the single-frame point cloud data to the theoretical surface of the polishing tool. If this distance is less than a preset distance threshold d_threshold (for example, d_threshold = 0.1 mm), then this point is considered a contact point. Mark all the points that meet the force and distance conditions as contact points, and the remaining points as non-contact points. The point cloud composed of all contact points is used as the contact state point cloud data.

[0122] Preferably, the recognition of the polishing contact state for the synchronous force-point cloud data includes:

[0123] Extract the origin of the sensor Cartesian coordinate system based on the polishing acting moment data, calculate the pose of the polishing head, and generate the polishing head pose data;

[0124] Calculate the distance between the polishing head and the workpiece according to the polishing head pose data to obtain the distance change data;

[0125] Calculate the moment based on the polishing acting moment data to obtain the force component value in the Z-axis direction and the moment component values around the X and Y axes respectively;

[0126] Calculate the force change rate based on the force component value in the Z-axis direction to generate the Z-axis force change rate;

[0127] When the distance change data is less than 0.05 mm, the force component value in the Z-axis direction is greater than 0.5 N, the square root of the sum of the squares of the moment component values around the X and Y axes is greater than 0.05 Nm, and the Z-axis force change rate is less than 0.2 N / s, mark the synchronous force-point cloud data as contact point cloud data;

[0128] When the distance change data is greater than 0.1 mm, the force component value in the Z-axis direction is less than 0.1 N, and the square root of the sum of the squares of the moment component values around the X and Y axes is less than 0.01 Nm, mark the synchronous force-point cloud data as off-contact point cloud data; otherwise, mark the synchronous force-point cloud data as critical contact point cloud data;

[0129] Integrate the point cloud states of the contact point cloud data, off-contact point cloud data, and critical contact point cloud data to obtain the contact state point cloud data.

[0130] In the embodiments of the present invention, the installation position and orientation of a six-axis force sensor (e.g., ATI Gamma) are determined. In the robot teach pendant, the coordinate transformation matrix \(T_{sensor\_flange}\) of the force sensor in the robot end flange coordinate system is recorded. Through the forward kinematics of the robot, the transformation matrix \(T_{flange\_base}\) of the robot end flange coordinate system in the base coordinate system is calculated. Using these two transformation matrices, the position of the origin of the sensor coordinate system in the robot base coordinate system is calculated as \(P_{sensor\_base}=T_{flange\_base}\times T_{sensor\_flange}\times P_{sensor\_local}\), where \(P_{sensor\_local}\) is the position of the origin of the sensor coordinate system in its own coordinate system (usually \([0,0,0]\)). Then, according to the force data \((F_x,F_y,F_z)\) and torque data \((M_x,M_y,M_z)\) measured by the force sensor, the pose of the polishing head is calculated. The polishing head is spherical, and the center of the sphere coincides with the origin of the sensor coordinate system. The rotation matrix \(R_{tool\_sensor}\) of the polishing head coordinate system relative to the sensor coordinate system is calculated using the torque data. The specific method is as follows: normalize the torque vector \((M_x,M_y,M_z)\) to obtain the rotation axis vector; calculate the rotation angle, which is the norm of the torque vector; according to the Rodriguez formula, calculate the rotation matrix \(R_{tool\_sensor}\). Finally, combining the position of the polishing head in the sensor coordinate system (i.e., the center of the sphere position), the pose data of the polishing head in the robot base coordinate system, including position and orientation, is calculated. An initial point set is selected on the surface fitting curve model of the workpiece, and this point set is evenly distributed within the projection area of the polishing head. Then, for each point on the polishing head (e.g., the uniformly distributed sampling points on the surface of the polishing head), the point closest to it is found in the initial point set. Calculate the distances between all these point pairs and take the average value as the distance for this iteration. Update the initial point set: for each point in the initial point set, find the point on the polishing head closest to it and use this point as the new initial point. Repeat the above steps until the distance change is less than a preset threshold (e.g., 0.001 mm) or the maximum number of iterations (e.g., 10 times) is reached. Take the finally obtained distance as the distance \(d\) between the polishing head and the workpiece. Calculate the difference in distance between two adjacent trigger moments as the distance change data \(\Delta d\). Read the force data \((F_x,F_y,F_z)\) and torque data \((M_x,M_y,M_z)\) measured by the six-axis force sensor from the synchronized force-point cloud data. Among them, \(F_z\) is the value of the force component in the Z-axis direction. \(M_x\) and \(M_y\) are the torque component values around the X-axis and Y-axis respectively. Numerically differentiate the value of the force component \(F_z\) in the Z-axis direction to calculate its change rate. The central difference method is used: for the force \(F_z(t)\) in the Z-axis direction at the current moment \(t\), its change rate \(dF_z / dt(t)=(F_z(t + \Delta t)-F_z(t-\Delta t)) / (2\Delta t)\), where \(\Delta t\) is the time interval between two adjacent trigger moments.Based on the obtained distance change data Δd, the Z-axis direction force component value Fz, the moment component values Mx and My about the X-axis and Y-axis, and the Z-axis force change rate dFz / dt, the contact state is judged. If Δd < 0.05 mm and Fz > 0.5 N, and dFz / dt < 0.2 N / s, then mark the synchronous force-point cloud data as contact point cloud data. These conditions indicate that a stable contact is maintained between the polishing head and the workpiece, and the contact force changes smoothly. Based on the obtained distance change data Δd, the Z-axis direction force component value Fz, the moment component values Mx and My about the X-axis and Y-axis, the disengagement contact state is judged. If Δd > 0.1 mm, Fz < 0.1 N, and then mark the synchronous force-point cloud data as disengagement contact point cloud data. These conditions indicate that the polishing head and the workpiece have been separated, and there is no obvious contact force or moment. If the synchronous force-point cloud data does not meet the conditions, it is marked as critical contact point cloud data. This indicates an uncertain contact state between the polishing head and the workpiece, about to disengage or with slight vibration. Combine the marked contact point cloud data, disengagement contact point cloud data, and critical contact point cloud data together to form a complete contact state point cloud data set.

[0131] Preferably, step S3 includes the following steps:

[0132] Step S31: Extract the contact state at the current moment according to the contact state point cloud data;

[0133] Step S32: Match the impedance parameters for the contact state at the current moment through a preset impedance adjustment rule to obtain real-time impedance adjustment parameters;

[0134] Step S33: Analyze the surface gradient at the point at the current moment through the surface fitting surface model, and then perform surface normal vector processing to obtain the real-time surface normal vector;

[0135] Step S34: Use the real-time surface normal vector to perform current normal pose analysis on the contact state point cloud data to generate current normal pose data;

[0136] Step S35: Analyze the grinding desired pose error of the surface fitting surface model through the current normal pose data to obtain the grinding normal pose error;

[0137] Step S36: Perform normal force / tangential force impedance error compensation according to the grinding normal pose error and the real-time impedance adjustment parameters, and perform inverse solution processing of the Jacobian matrix to obtain the reverse impedance compensation amount.

[0138] In an embodiment of the present invention, if the point cloud data at the current moment is marked as "contact point cloud data", the "contact" state is extracted as the contact state at the current moment. If the point cloud data at the current moment is marked as "disengaged contact point cloud data", the "disengaged contact" state is extracted as the contact state at the current moment. If the point cloud data at the current moment is marked as "critical contact point cloud data", the "critical contact" state is extracted as the contact state at the current moment. An impedance parameter adjustment rule library is established. The rule library contains the corresponding impedance parameters under different contact states. For example: Contact state: Set a relatively high stiffness K (e.g., K = 1500 N / m) and a moderate damping B (e.g., B = 120 Ns / m) to ensure the polishing quality and stability. Disengaged contact state: Set a relatively low stiffness K (e.g., K = 500 N / m) and a small damping B (e.g., B = 50 Ns / m) to reduce the impact of the robot movement and quickly respond to contact. Critical contact state: Set a medium stiffness K (e.g., K = 1000 N / m) and a large damping B (e.g., B = 150 Ns / m) to suppress the occurring vibration. According to the contact state at the current moment, the corresponding impedance parameters (K, B, M) are selected from the rule library. According to the coordinates (x, y, z) of the polishing point at the current moment, the corresponding parameters (u, v) are found on the B-spline surface model. Then, the partial derivatives of the surface at this parameter point are calculated and where P(u, v) represents a point on the surface. These two partial derivative vectors respectively represent the tangent vectors of the surface in the u direction and the v direction. The cross product of these two tangent vectors is taken to obtain the normal vector at this point Normalize the normal vector N to obtain the unit normal vector n = N / ||N||. This unit normal vector n is the real-time surface normal vector. For each point Pi in the contact state point cloud data, calculate its distance di to the polishing head. If di is less than a preset threshold (e.g., 0.1 mm), then this point is considered an effective contact point. Calculate the average position of all effective contact points. Take the real-time surface normal vector n as the desired normal vector of the polishing head at the current moment. Combine the average position and the desired normal vector to form the current normal pose data. Find the nearest control point on the surface model according to the average position and calculate the normal vector at this point. Calculate the angle θ between the desired normal vector of the polishing head at the current moment and the normal vector of the nearest control point as the normal angle error. Calculate the distance d between the average position and the polishing position point at the current moment as the normal position error. Combine the normal angle error θ and the normal position error d to form the grinding normal pose error. Convert the normal angle error θ to the desired torque compensation amount ΔM = K_θ × θ, where K_θ is the angle stiffness coefficient (e.g., K_θ = 10 Nm / rad). Convert the normal position error d to the desired force compensation amount ΔF = K × d, where K is the stiffness selected in step S32. Since impedance control mainly focuses on the normal force, the normal force error is mainly compensated here. Then calculate the impedance torque of the robot end in the Cartesian space: F_i = [0, 0, ΔF, ΔMx, ΔMy, 0], where the tangential force is expected to be 0, and ΔMx and ΔMy are obtained by decomposing ΔM around the x and y axes. Then, use the Jacobian matrix J of the robot to convert the impedance torque in the Cartesian space to the compensation amount Δq = J^T × F_i in the joint space, where T represents the matrix transpose. When J^T is not a square matrix, it can be solved by using the pseudo-inverse or the damped least squares method. The obtained Δq is the reverse impedance compensation amount, and this compensation amount will be superimposed on the joint control command of the robot to achieve precise control of the normal force and posture.

[0139] Preferably, step S34 includes the following steps:

[0140] Step S341: Extract the three-dimensional coordinates of the current polishing point according to the contact state point cloud data;

[0141] Step S342: Take the real-time surface normal vector as the direction vector and use the three-dimensional coordinates of the current polishing point as the origin to construct the normal axis at the current moment;

[0142] Step S343: Calculate the perpendicular projection point of the three-dimensional coordinates of the current polishing point through the normal axis at the current moment to obtain the current normal position;

[0143] Step S344: Extract the three-dimensional coordinates of the polishing point at the previous moment according to the contact state point cloud data;

[0144] Step S345: Calculate the perpendicular projection point of the three-dimensional coordinates of the polishing point at the previous moment through the normal axis at the current moment to obtain the normal position at the previous moment;

[0145] Step S346: Calculate the normal velocity based on the current normal position and the normal position at the previous moment to obtain the current normal velocity;

[0146] Step S347: Combine the state variables according to the current normal position and the current normal velocity to generate the current normal pose data.

[0147] In an embodiment of the present invention, if the point cloud data at the current moment is marked as "contact point cloud data", the three-dimensional coordinates (x, y, z) of all points are extracted therefrom. If the point cloud at the current moment contains multiple points, the centroid of all these points is calculated, and the centroid coordinates are used as the three-dimensional coordinates (x_c, y_c, z_c) of the current polishing point. The calculation formula is: x_c = (Σxi) / n, y_c = (Σyi) / n, z_c = (Σzi) / n, where n is the number of points in the point cloud, and xi, yi, zi are the coordinates of each point. If the point cloud data is empty (for example, in a state of out-of-contact), the coordinates of the polishing point at the previous contact moment are used. Based on the real-time surface normal vector n = (nx, ny, nz) obtained in step S33 and the three-dimensional coordinates P_c = (x_c, y_c, z_c) of the current polishing point obtained in step S341, a straight line passing through point P_c and with the direction of n is constructed as the normal axis at the current moment. The parametric equation of this normal axis can be expressed as: L(t) = P_c + t×n, where t is a parameter. This straight line represents the normal direction of the workpiece surface at the current polishing point. The current polishing point P_c = (x_c, y_c, z_c) is perpendicularly projected onto the normal axis L(t) constructed in step S342. Therefore, the coordinates of the projection point are the same as the origin. However, for the sake of maintaining the consistency of the process, the calculation process is still carried out: substituting the projection time (t_proj) into the parametric equation of the normal axis, the projection point P_proj_c = P_c + t_proj×n = P_c is obtained. This projection point P_proj_c is the current normal position. From the contact state point cloud data, the point cloud data corresponding to the previous trigger moment (i.e., the previous scanning period) is extracted. If the point cloud data at the previous moment is marked as "contact point cloud data", the three-dimensional coordinates of all points are extracted therefrom. If the point cloud at the previous moment contains multiple points, the centroid of all these points is calculated, and the centroid coordinates are used as the three-dimensional coordinates (x_p, y_p, z_p) of the polishing point at the previous moment. If the point cloud data at the previous moment is empty or marked as "out-of-contact point cloud data", the coordinates of the polishing point at an earlier contact moment are used until a valid contact point coordinate is found. The polishing point P_p = (x_p, y_p, z_p) at the previous moment is perpendicularly projected onto the normal axis L(t) of the current moment constructed in step S342. Substituting t_proj into the parametric equation of the normal axis, the projection point P_proj_p = P_c + t_proj×n is obtained. This projection point P_proj_p is the normal position of the previous moment relative to the normal axis of the current moment. Based on the current normal position obtained in step S343 and the normal position of the previous moment obtained in step S345, the normal velocity is calculated. Since both of these two points are located on the normal axis of the current moment, the distance between them can be directly calculated by the coordinate difference. Calculate the distance vector d = P_proj_c - P_proj_p between the two projection points. Calculate the time interval Δt between the two trigger moments.The normal velocity v_n = (d·n) / Δt. Here, (d·n) represents the projection of the vector d on the normal vector n, that is, the change in the normal distance. Combining the current normal position P_proj_c (expressed as three-dimensional coordinates) obtained in step S343 and the current normal velocity v_n (a scalar) obtained in step S346 forms the current normal pose data. This data can be represented as a vector: [P_proj_c_x, P_proj_c_y, P_proj_c_z, v_n]. This vector contains the position and velocity information of the polishing head along the normal direction of the workpiece surface at the current moment.

[0148] As an example of the present invention, refer to Figure 3 shown in Figure 1 the detailed implementation step flow diagram of step S4 in

[0149] Step S41: Obtain the current robot joint angle values;

[0150] In the embodiment of the present invention, a real-time controller (for example, a controller based on the EtherCAT bus with a cycle period of 1 ms) is used to establish communication with the driver of an industrial robot (for example, ABB IRB 1200). Through the EtherCAT bus, the real-time controller sends an instruction to each joint driver to read the joint angle. The encoder inside the driver (for example, an absolute encoder with a resolution of 16 bits) precisely measures the current rotation angle of each joint motor. The driver converts the angle value read by the encoder into a joint angle value (unit: radian) and feeds it back to the real-time controller through the EtherCAT bus. The real-time controller receives the joint angle values fed back by all joint drivers and forms a vector q_current = [q1, q2, q3, q4, q5, q6] containing the current angles of all joints, where qi represents the current angle of the i-th joint. To ensure the reliability of the data, CRC check is used to check the received data. If the check fails, the read instruction is resent until the correct data is received. The frequency of obtaining the joint angle values is consistent with the cycle period of the controller, that is, once every 1 ms.

[0151] Step S42: Perform joint space compensation amount conversion on the current robot joint angle values through the inverse impedance compensation amount to obtain the joint space compensation amount;

[0152] In the embodiment of the present invention, the received reverse impedance compensation amount Δq = [Δq1, Δq2, Δq3, Δq4, Δq5, Δq6]. Since Δq is already the compensation amount in the joint space, this step is mainly for data type conversion and unit unification. Convert the data type of Δq from floating point to the same data type as the joint angle value (for example, double-precision floating point). Ensure that the unit of Δq is consistent with the unit of the joint angle value (both are radians). If the units are inconsistent, perform the corresponding conversion. For example, if the unit of Δq is degrees, it needs to be multiplied by π / 180 to convert to radians. Since the reverse impedance compensation amount Δq is already in the joint space and no additional conversion is required, for the integrity of the embodiment, assign the reverse impedance compensation amount Δq to the joint space compensation δq, δq = Δq. Finally, δq is directly used as the compensation amount for the current robot joint angle.

[0153] Step S43: Correct the spatial joint angle according to the joint space compensation amount, and synthesize joint control instructions to obtain a joint drive control signal;

[0154] In the embodiment of the present invention, based on the obtained current robot joint angle value q_current and the joint space compensation amount δq obtained in step S42, calculate the corrected joint angle q_corrected. The calculation formula is: q_corrected = q_current + δq. Limit each element of q_corrected within the motion range of the corresponding robot joint. If the corrected angle of a certain joint exceeds its motion range, truncate it to the boundary value. For example, if q_corrected[i] > q_max[i], then q_corrected[i] = q_max[i]; if q_corrected[i] < q_min[i], then q_corrected[i] = q_min[i], where q_max[i] and q_min[i] respectively represent the upper and lower limits of the angle of the i-th joint. Convert the corrected joint angle q_corrected into a control instruction format recognizable by the joint driver. For example, convert the angle value into encoder count values. The conversion formula is: counts = (q_corrected / (2π)) × encoder_resolution, where encoder_resolution is the resolution of the encoder. Combine the encoder count values of each joint into a complete joint control instruction data packet. The format of the data packet follows the requirements of the EtherCAT communication protocol and includes fields such as a synchronization header, data length, data type, joint number, encoder count value, checksum, etc. This data packet is the joint drive control signal.

[0155] Step S44: Send the joint drive control signal to the driver through the communication bus to control the correction of each joint angle of the polishing industrial robot, so as to achieve the reverse impedance control of the automated polishing equipment.

[0156] In the embodiment of the present invention, the real-time controller sends the joint drive control signal generated in step S43 to each joint driver of the industrial robot through the EtherCAT bus. The EtherCAT bus adopts the master-slave communication mode, with the real-time controller as the master station and each joint driver as the slave station. The controller broadcasts and sends data packets to all slave stations according to a fixed cycle period (for example, 1 ms). Each slave station receives its own control instruction according to the joint number in the data packet. After receiving the control instruction, the driver performs CRC check to ensure the correctness of the data. After the check passes, the PID controller inside the driver calculates the motor control voltage according to the received encoder count value (desired joint angle) and the actual encoder feedback value (actual joint angle). The driver drives the joint motor to rotate by controlling the motor voltage, so that the joint angle approaches the desired value. Through this closed-loop control method, precise control of each joint angle is achieved, thereby realizing the reverse impedance control of the entire industrial robot, so that the end effector contacts and polishes the workpiece according to the desired force and pose. The whole process runs at a high speed and stably under the control of the real-time controller with a cycle period of 1 ms.

[0157] Preferably, the present invention also provides an automated equipment control system that executes the automated equipment control method as described above. The automated equipment control system includes:

[0158] A polishing instruction generation module, which is used to scan the surface of the target polishing part by using a high-frequency laser scanner to obtain the polished surface point cloud data; perform B-spline surface fitting according to the polished surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control according to the surface fitting surface model to obtain a polishing control instruction;

[0159] A polishing state monitoring module, which is used to control the polishing operation of the polishing industrial robot based on the polishing control instruction, and perform pulse trigger monitoring to generate a synchronous trigger pulse sequence; perform polishing contact state detection according to the synchronous trigger pulse sequence to generate contact state point cloud data;

[0160] An impedance error analysis module, which is used to perform grinding desired pose error analysis based on the contact state point cloud data and perform impedance error compensation to obtain a reverse impedance compensation amount;

[0161] An equipment joint impedance control module, which is used to correct the joint angle of the polishing industrial robot through the reverse impedance compensation amount to achieve the reverse impedance control of the automated polishing equipment.

[0162] Preferably, the present invention further provides a computer-readable storage medium storing a computer program, which when executed implements the automated equipment control method described in any one of the above.

[0163] This application lies in that, through high-frequency laser scanners to perform high-precision scanning on the surface of the target polished part, combined with B-spline surface fitting technology to generate a surface fitting surface model, it can accurately construct the geometric shape of complex curved surface parts, effectively eliminating the deviation caused by part manufacturing errors and clamping deformations. On this basis, the generated polishing control instructions not only contain position information, but also incorporate refined control parameters based on the surface model, ensuring a high degree of consistency between the initial polishing path and the part surface, reducing the risk of excessive grinding in the initial stage, and significantly improving the polishing accuracy and surface finish. By using a synchronous trigger pulse sequence, the coordinated operation of the high-frequency laser scanner and the six-dimensional force sensor is realized, enabling real-time and accurate monitoring of the contact state between the grinding wheel and the part surface during the polishing process. By analyzing the synchronous force-point cloud data, the system can accurately judge key information such as the magnitude, direction, and contact area of the grinding force, and dynamically adjust the polishing parameters according to the actual grinding situation, effectively avoiding the problem of grinding force fluctuations caused by the inability to real-time sense the contact state in traditional methods, and significantly improving the stability and consistency of the polishing process. Through the analysis of the grinding desired pose error and the reverse impedance error compensation, the joint angle correction of the polishing industrial robot is realized. This technology can perform fine-tuning within a certain range according to the magnitude and direction of the actual grinding force to maintain the constancy of the grinding force. Even when facing variable curvature surfaces, the robot can adaptively adjust its own posture and motion trajectory to ensure that the grinding wheel always contacts the part surface with an appropriate force. This reverse impedance control method effectively solves the problem of grinding force fluctuations caused by fixed trajectories and the inability to cope with curvature changes in traditional position control, significantly improving the polishing quality and production efficiency of complex curved surface parts.

[0164] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, it is intended to encompass all changes falling within the meaning and scope of the equivalent elements of the application documents within the present invention.

[0165] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather will conform to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for controlling an automated device, characterized in that: Applied to automated polishing equipment, the automated polishing equipment includes a polishing industrial robot, a six-dimensional torque sensor, a vibration sensor, a high-frequency laser scanner and a control unit, the control unit includes a real-time controller and a driver, the six-dimensional torque sensor, the vibration sensor and the high-frequency laser scanner are electrically connected to the real-time controller, and the polishing industrial robot is electrically connected to the real-time controller through the driver; the automated equipment control method includes the following steps: Step S1: Scan the surface of the target polishing part with a high-frequency laser scanner to obtain polishing surface point cloud data; perform B-spline surface fitting according to the polishing surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control according to the surface fitting surface model to obtain a polishing control instruction; Step S2: Based on the polishing control instruction, the polishing industrial robot is controlled to perform the polishing operation, and pulse trigger monitoring is performed to generate a synchronous trigger pulse sequence; polishing contact state detection is performed according to the synchronous trigger pulse sequence to generate contact state point cloud data; Step S3: performing grinding desired posture error analysis based on the contact state point cloud data, and performing reverse impedance error compensation on the polishing industrial robot to obtain a reverse impedance compensation amount; Step S4: Correcting the joint angle of the polishing industrial robot by using the reverse impedance compensation amount to achieve reverse impedance control of the automated polishing equipment.

2. The automation equipment control method according to claim 1, characterized in that: The preliminary polishing control according to the surface fitting surface model in step S1 includes: Evaluate the surface roughness of the machining area based on the surface fitting surface model; Obtaining polishing part material property data; calculating the grinding force of the surface roughness of the processing area based on the initial grinding force range and the polishing part material property data to generate regional surface grinding force data, wherein the initial grinding force range is 5N-10N; Surface regional curvature is extracted according to the surface fitting surface model to generate regional curvature distribution data; According to the surface roughness of the processing area and the regional curvature distribution data, the polishing control points are set at 2-5 points per square millimeter; Perform contour grinding path planning for the polishing control points, with the contour interval being 0.2 mm to 0.5 mm, the fillet radius at the path corner being not less than 0.3 mm, the straight-line distance between the path start point and the end point being not less than 10 mm, and the total path length error being no more than ±2.0 mm, so as to obtain grinding path data; Preliminary polishing control instruction encoding is performed according to the regional surface grinding force data and the grinding path data to obtain the polishing control instruction.

3. The automation equipment control method according to claim 1, characterized in that: Controlling the polishing operation of the polishing industrial robot based on the polishing control instruction and performing pulse triggering monitoring in step S2 includes: The real-time controller is used to analyze the polishing control instructions, and then the polishing industrial robot is controlled by the driver to perform the polishing operation, and the vibration sensor is used to monitor the joint space vibration signal of the polishing industrial robot in real time; Extracting vibration signal features from the joint space vibration signal to obtain joint vibration feature data; Perform joint vibration feature aggregation based on joint vibration feature data, perform vibration cycle recognition, and generate joint vibration cycles; Based on the preset polishing scan threshold, the vibration cycle scan trigger times are counted through the joint vibration cycle to obtain the single cycle trigger times; The relative time delay of each trigger point is calculated according to the single-cycle triggering times and the joint vibration characteristic data to obtain the triggering delay sequence; Polishing stability state is evaluated based on joint vibration characteristic data to obtain a polishing stability index; Acquire real-time clock data; perform adaptive monitoring of trigger pulse processing on the trigger delay sequence based on the polishing stability index and the real-time clock data to generate a synchronous trigger pulse sequence.

4. The automation equipment control method according to claim 1, characterized in that: The polishing contact state detection according to the synchronous trigger pulse sequence in step S2 includes: Based on the synchronous trigger pulse sequence, the high-frequency laser scanner is controlled to perform triggered polishing surface scanning to generate a high-frequency scanning original point cloud; According to the synchronous trigger pulse sequence, the six-dimensional torque sensor is controlled to measure the polishing torque of the end effector of the polishing industrial robot to generate polishing torque data; Data filtering and calibration are performed based on the high-frequency scanning original point cloud, and point cloud segmentation is performed to obtain single-frame point cloud data; Time stamp synchronization of polishing action torque data is performed through single-frame point cloud data to generate synchronized force-point cloud data; Polishing contact state identification is performed on the synchronous force-point cloud data to generate contact state point cloud data.

5. The automation equipment control method according to claim 4, characterized in that: The polishing contact state identification of the synchronous force-point cloud data includes: The origin of the sensor Cartesian coordinate system is extracted according to the polishing torque data, and the polishing head posture is calculated to generate the polishing head posture data; Calculate the distance between the polishing head and the workpiece according to the polishing head posture data to obtain distance change data; The torque is calculated according to the polishing torque data, and the force component value in the Z-axis direction and the torque component values ​​around the X and Y axes are obtained respectively; The force change rate is calculated according to the force component value in the Z-axis direction to generate the Z-axis force change rate; When the distance change data is less than 0.05 mm, the Z-axis force component value is greater than 0.5 N, the square root of the sum of the squares of the moment component values ​​around the X and Y axes is greater than 0.05 Nm, and the Z-axis force change rate is less than 0.2 N / s, the synchronous force-point cloud data is marked as contact point cloud data; When the distance change data is greater than 0.1mm, the Z-axis force component value is less than 0.1N, and the square root of the sum of the squares of the moment component values ​​around the X and Y axes is less than 0.01Nm, the synchronous force-point cloud data is marked as out-of-contact point cloud data; otherwise, the synchronous force-point cloud data is marked as critical contact point cloud data; The contact point cloud data, the disengagement point cloud data and the critical contact point cloud data are integrated into point cloud states to obtain the contact state point cloud data.

6. The automation equipment control method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: extracting the current contact state according to the contact state point cloud data; Step S32: performing impedance parameter matching on the contact state at the current moment according to a preset impedance adjustment rule to obtain a real-time impedance adjustment parameter; Step S33: analyzing the surface gradient at the current time point through the surface fitting surface model, and then performing surface normal vector processing to obtain a real-time surface normal vector; Step S34: using the real-time surface normal vector to perform current normal posture analysis on the contact state point cloud data to generate current normal posture data; Step S35: performing grinding expected posture error analysis on the surface fitting surface model through the current normal posture data to obtain the grinding normal posture error; Step S36: Perform normal force / tangential force impedance error compensation according to the grinding normal posture error and the real-time impedance adjustment parameters, and perform inverse solution of the Jacobian matrix to obtain the reverse impedance compensation amount.

7. The automation equipment control method according to claim 6, characterized in that: Step S34 includes the following steps: Step S341: extracting the three-dimensional coordinates of the current polishing point according to the contact state point cloud data; Step S342: using the real-time surface normal vector as the direction vector and the three-dimensional coordinates of the current polishing point as the origin to construct the normal axis at the current moment; Step S343: Calculate the vertical projection point of the three-dimensional coordinates of the current polishing point through the normal axis at the current moment to obtain the current normal position; Step S344: extracting the three-dimensional coordinates of the polishing point at the previous moment according to the contact state point cloud data; Step S345: Calculate the vertical projection point of the three-dimensional coordinates of the polishing point at the previous moment through the normal axis at the current moment to obtain the normal position at the previous moment; Step S346: Calculate the normal velocity based on the current normal position and the normal position at the previous moment to obtain the current normal velocity; Step S347: Combine the state quantities according to the current normal position and the current normal velocity to generate the current normal posture data.

8. The automation equipment control method according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Obtain the current robot joint angle value; Step S42: converting the current robot joint angle value into a joint space compensation amount by using the reverse impedance compensation amount to obtain the joint space compensation amount; Step S43: correcting the spatial joint angle according to the joint spatial compensation amount, and synthesizing the joint control instructions to obtain a joint drive control signal; Step S44: sending the joint drive control signal to the driver through the communication bus to control the angle correction of each joint of the polishing industrial robot to achieve reverse impedance control of the automated polishing equipment.

9. An automation equipment control system, characterized in that: Used to execute the automation equipment control method according to claim 1, the automation equipment control system comprises: The polishing instruction generation module is used to scan the surface of the target polishing part with a high-frequency laser scanner to obtain the polishing surface point cloud data; perform B-spline surface fitting according to the polishing surface point cloud data to generate a surface fitting surface model; perform preliminary polishing control according to the surface fitting surface model to obtain a polishing control instruction; The polishing state monitoring module is used to control the polishing operation of the polishing industrial robot based on the polishing control instruction, and perform pulse triggering monitoring to generate a synchronous trigger pulse sequence; perform polishing contact state detection according to the synchronous trigger pulse sequence to generate contact state point cloud data; The impedance error analysis module is used to analyze the grinding desired posture error based on the contact state point cloud data, and to perform impedance error compensation to obtain the reverse impedance compensation amount; The equipment joint impedance control module is used to correct the joint angle of the polishing industrial robot through the reverse impedance compensation amount to realize the reverse impedance control of the automated polishing equipment.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed, the automation equipment control method according to any one of claims 1 to 8 is implemented.

Citation Information

Cited By

  • Automatic control method for milling cutter machining

    CN120595714A

  • Wafer thinning device and thinning method

    CN121018325A

  • Surface treatment method and device after machining of automobile crankshaft

    CN121132422A

  • Automobile crankshaft machining post-surface treatment method and device

    CN121132422B

  • Full-band error cooperative control space-time co-fusion polishing method and device

    CN121715922A