A dynamic monitoring and optimization method for surface film flatness of a body robot

Abstract

The application discloses a kind of curved surface film covering flatness dynamic monitoring and optimization method of embodied robot, it is related to flexible electronic manufacturing technical field, and this method includes: coarse adjustment step, by projection device projection interference fringe to flexible film material surface, by first camera image acquisition and calculate first height distribution and global flatness index, by first controller driving clamp is adjusted, so that index satisfies first threshold value;Fine adjustment step, by stereo vision system obtains film material three-dimensional topography and obtains second height distribution, compares it with target topography and obtains residual error, based on residual error and pre-calibrated clamp-topography sensitivity matrix, by the optimization model with regularization calculation each clamp fine adjustment amount and drive execution, so that flatness index satisfies more strict second threshold value;Coarse adjustment and fine adjustment are cyclically executed, and control weight is dynamically adjusted to realize smooth transition and closed-loop control.The application realizes the high-precision, adaptive flatness control of film covering on complex curved surface.

Description

Technical Field

[0001] This invention relates to the field of flexible electronics manufacturing technology, and in particular to a method for dynamic monitoring and optimization of the flatness of curved surface coating on a robotic body. Background Technology

[0002] Flexible electronic products, due to their ultra-thin and flexible characteristics, have shown broad application prospects in the field of intelligent surface sensing for complex curved surface equipment such as unibody robots. However, in the manufacturing process of bonding flexible electronic films to the complex curved surfaces of robots, various defects such as local unevenness, conductor breakage, and interface delamination are easily generated due to changes in surface curvature and uneven distribution of bonding stress, which seriously affect the structural integrity and functional reliability of the product.

[0003] Existing lamination process monitoring technologies primarily rely on traditional tensile testing devices. These devices are mostly static, single-point measurements, unable to effectively capture transient, multi-dimensional dynamic deformations caused by factors such as robot motion, abrupt changes in surface geometry, or environmental thermal disturbances during the lamination of complex curved surfaces. Therefore, monitoring results based on limited static point data cannot fully characterize the three-dimensional mechanical state of the entire lamination process, leading to lag in manufacturing system response, inaccurate control, large fluctuations in product quality, and high rework rates, significantly limiting the application of flexible electronic products in high-end equipment.

[0004] In the prior art, there are also some devices that attempt to improve upon this technology. For example, Chinese utility model patent CN218557941U discloses a film stretching device that uses a lead screw to drive roller assemblies away from each other to stretch and wind the film, keeping it in a taut state. However, this device only works under static conditions. When the film undergoes dynamic deformation due to heat during processes such as laser welding, its static single-point control mode cannot achieve real-time characterization and fine-tuning of the three-dimensional morphology, making it difficult to ensure a continuous and tight fit between the film and the curved surface of the part.

[0005] For example, Chinese invention patent CN119261173A discloses another film stretching device, which achieves rapid control of the film through a motor and a movable positioning mechanism. However, this device lacks a monitoring mechanism for the real-time state of the film and fails to form a closed-loop control system with the film morphology information and the adjustment device. Therefore, it cannot achieve precise and adaptive control of the film bonding state.

[0006] In summary, existing technologies cannot meet the high-precision monitoring and real-time control requirements for dynamic, three-dimensional, and full-field flatness during the complex curved surface lamination process of embodied robots. Therefore, there is an urgent need in this field for a novel monitoring and optimization method that can monitor the dynamic deformation of flexible electronic products in real time and accurately, and form a rapid closed-loop feedback with the actuator, in order to improve the flexibility, response speed, and product yield of the manufacturing process. Summary of the Invention

[0007] To address this issue, this invention provides a method for dynamic monitoring and optimization of the flatness of curved surface lamination using a robotic body. This method solves the problem that the static, single-point monitoring method in the prior art cannot respond in real time to the dynamic three-dimensional deformation during the lamination process, resulting in poor bonding accuracy and high defect rate.

[0008] To address the aforementioned technical problems, embodiments of the present invention provide a method for dynamic monitoring and optimization of the flatness of curved surface coating on a unibody robot, the method comprising:

[0009] Coarse adjustment steps: Interference fringes are projected onto the surface of the flexible membrane material fixed on the fixture using a projection device, and the fringe image is acquired by a first camera; the first height distribution of the membrane surface is calculated based on the phase change of the fringe image, and the global flatness index is calculated accordingly; based on the global flatness index, a global adjustment command is generated by a first controller to drive the fixture to perform pose adjustment so that the global flatness index meets the first threshold.

[0010] Fine-tuning steps: The three-dimensional morphology of the flexible membrane surface is obtained through a stereo vision system to obtain a second height distribution; the second height distribution is compared with the target morphology to obtain a residual; based on the residual and a pre-calibrated fixture-morphology sensitivity matrix, the fine-tuning amount of each fixture is calculated through a regularized optimization model, and a second control command is generated to drive the fixture to execute the fine-tuning amount so that the flatness index corresponding to the second height distribution meets the second threshold.

[0011] Closed-loop control steps: Throughout the coating process, the coarse adjustment step and the fine adjustment step are executed cyclically, and the control weights of the two steps are dynamically adjusted according to the iteration process to achieve a smooth transition from coarse adjustment to fine adjustment and dynamic closed-loop control.

[0012] Preferably, calculating the first height distribution of the film surface based on the phase change of the stripe image, and calculating the global smoothness index accordingly, specifically includes:

[0013] The phase distribution is obtained from the stripe image using a phase extraction and unpacking algorithm. And using calibration coefficients Through formula The change in height distribution was calculated. The global flatness index includes the root mean square error of the height distribution. Flatness error and maximum gradient ;

[0014] in

[0015] ;

[0016] ;

[0017] ;

[0018] in, The measurement area; For pixel coordinates The height value measured at the location; For the entire measurement area All height values The arithmetic mean, i.e. , Indicates the measurement area The total number of pixels within the region, i.e., the area of ​​the region; This indicates the deviation of the height of a single pixel from the average height. Indicates the measurement area The maximum value among all height values ​​within; Indicates the measurement area The minimum value among all height values; The gradient operator for the height function, .

[0019] Preferably, in the coarse adjustment step, when the Less than or equal to the first threshold and When the price drops several times in a row, the process is considered to proceed to the fine-tuning step.

[0020] Preferably, in the coarse adjustment step, the first controller is a PID controller, and the global adjustment command generated by it includes a global pose correction amount. The global pose correction amount The calculation formula is:

[0021] ;

[0022] in, , , These are proportional gain, integral gain, and derivative gain, respectively. To control deviation; This is the cumulative sum of the deviations; The duration of the cycle; This is the deviation value from the previous control cycle.

[0023] Preferably, in the fine-tuning step, the mathematical expression of the optimization model is:

[0024] ;

[0025] in, For the fine-tuning amount of the fixture; For residuals; This is the fixture-topography sensitivity matrix; This is the weight matrix; and The regularization coefficient is used. For smoothing operators; Let L2 norm be represented; its analytical solution is:

[0026] ;

[0027] in, This indicates transpose.

[0028] Preferably, in the fine-tuning step, when the root mean square error... Less than the second threshold and And the single-step adjustment amount of the fixture The fine-tuning steps were determined to have converged, among which... This is a preset positive threshold.

[0029] Preferably, in the closed-loop control step, a weighted scheduling function is used to smoothly switch between the coarse-tuning step and the fine-tuning step. The weighted scheduling function is:

[0030] Coarse weighting Fine-tuning the weights ,in Let κ be the number of iterations and κ be the decay constant; the discrete execution ratio is: number of iterations. At that time, the coarse adjustment weight : fine adjustment weight = 0.80 : 0.20; At that time, the coarse adjustment weight and the fine adjustment weight were 0.50:0.50. At that time, the coarse adjustment weight and the fine adjustment weight were 0.10:0.90.

[0031] Preferably, the method further includes an execution constraint step, which limits the amplitude and rate of the global adjustment command and the fine-tuning amount, wherein the single-step adjustment amount satisfies Adjust the rate to meet And the stroke and rotation angle of the fixture are limited, among which... For the fine-tuning amount of the fixture, This refers to the adjustment rate of the fixture.

[0032] This invention also provides a dynamic monitoring and optimization system for the flatness of curved surface coating on a unibody robot. This system is used to implement the aforementioned method for dynamic monitoring and optimization of the flatness of curved surface coating on a unibody robot, specifically including:

[0033] The optical monitoring subsystem includes a projection device for projecting interference fringes, a first camera for acquiring fringe images, and a stereo vision system for three-dimensional topography reconstruction.

[0034] The multi-degree-of-freedom fixture subsystem includes an electric three-dimensional assembly stage and multiple independently controlled fixtures mounted thereon for fixing and adjusting the pose of the flexible membrane material.

[0035] The control subsystem, which is communicatively connected to the optical monitoring subsystem and the multi-degree-of-freedom fixture subsystem, is configured to execute the coarse adjustment step, the fine adjustment step, and the closed-loop control step.

[0036] This invention also provides a computer storage medium storing a computer software product, the computer software product including several instructions to cause a computer device to execute the above-described method for dynamic monitoring and optimization of the flatness of unibody robot curved surface coating.

[0037] As can be seen from the above technical solutions, this invention application has the following beneficial effects:

[0038] (1) This invention effectively solves the contradiction between accuracy and efficiency of a single monitoring method by combining coarse adjustment of interference fringes with fine adjustment of three-dimensional reconstruction. In the coarse adjustment stage, interference fringes are used to quickly identify and correct large-scale warping and wrinkles, rapidly improving the flatness to the initial threshold (e.g., 0.15 mm), greatly improving the initial convergence speed. In the fine adjustment stage, high-precision three-dimensional topography is obtained through stereo vision, and local micron-level corrections (e.g., 0.02 mm) are performed based on the sensitivity matrix and regularization optimization algorithm, ultimately achieving a high-precision flatness that cannot be achieved by traditional static methods. This strategy of first adjusting the overall and then the local significantly optimizes the efficiency of the adjustment process while ensuring the final quality.

[0039] (2) This invention does not simply superimpose two monitoring methods, but achieves deep information fusion and closed-loop control through algorithms. First, it innovatively introduces a weighted scheduling mechanism based on the number of iterations, realizing a smooth transition and adaptive fusion of coarse and fine adjustment control commands, avoiding abrupt changes and oscillations when switching control strategies. Second, the system deeply couples visual shape information with the mechanical adjustment capability of the fixture through a mathematical model (sensitivity matrix), and incorporates feedforward compensation (such as robot motion and thermal deformation prediction), forming a closed-loop system capable of real-time perception, intelligent decision-making, and precise execution. This enables the method to effectively cope with complex working conditions such as robot dynamic motion, sudden changes in surface curvature, and environmental thermal disturbances, maintaining the stability and high quality of the film application process. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Referring to the drawings will make the features and advantages of the present invention clearer. The drawings are illustrative and should not be construed as limiting the present invention in any way. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0041] Figure 1 This is a flowchart of a method for dynamic monitoring and optimization of the flatness of a curved surface coating on a holographic robot, provided by the present invention.

[0042] Figure 2 This is a front view of the hardware system of the method of the present invention;

[0043] Figure 3 This is a side view of the hardware system of the method of the present invention;

[0044] Figure 4 This is a partial sectional view of the side view of the hardware system of the method of the present invention.

[0045] Explanation of reference numerals in the accompanying drawings: 1. Fixed base plate; 2. Horizontal lead screw moving platform; 3. Horizontal slide table; 4. Base plate; 5. Stereo camera fixture; 6. Stereo camera; 7. Support device; 8. First servo motor; 9. Electric 3D combination stage; 10. Fixture; 11. Flexible electronic product; 12. Processed part; 13. Second servo motor; 14. Vertical slide table; 15. Laser profilometer; 16. Vertical lead screw moving platform; 17. Projection device; 18. Tripod. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1: To address the problem that existing static, single-point monitoring methods cannot respond in real time to the dynamic three-dimensional deformation during the lamination process, resulting in poor bonding accuracy and high defect rate, such as... Figure 1 As shown, this invention proposes a method for dynamic monitoring and optimization of the flatness of curved surface coating on a robotic body. The method includes:

[0048] Coarse adjustment steps: Interference fringes are projected onto the surface of the flexible membrane material fixed on the fixture using a projection device, and the fringe image is acquired by a first camera; the first height distribution of the membrane surface is calculated based on the phase change of the fringe image, and the global flatness index is calculated accordingly; based on the global flatness index, a global adjustment command is generated by a first controller to drive the fixture to perform pose adjustment so that the global flatness index meets the first threshold.

[0049] Fine-tuning steps: The three-dimensional morphology of the flexible membrane surface is obtained through a stereo vision system to obtain the second height distribution; the second height distribution is compared with the target morphology to obtain the residual; based on the residual and the pre-calibrated fixture-morphology sensitivity matrix, the fine-tuning amount of each fixture is calculated through a regularized optimization model, and a second control command is generated to drive the fixture to execute the fine-tuning amount so that the flatness index corresponding to the second height distribution meets the second threshold.

[0050] Closed-loop control steps: Throughout the coating process, coarse adjustment and fine adjustment steps are executed cyclically, and the control weights of the two steps are dynamically adjusted according to the iteration process to achieve a smooth transition from coarse adjustment to fine adjustment and dynamic closed-loop control.

[0051] As can be seen from the above technical solution, this invention proposes a method for dynamic monitoring and optimization of the flatness of embodied robot curved surface coating. By constructing a hierarchical dynamic control system of coarse adjustment, fine adjustment, and closed-loop control, high-precision adaptive bonding of flexible membrane materials on complex curved surfaces is achieved. In the coarse adjustment step, the system uses projection interference fringes and phase analysis technology to quickly obtain the overall flatness of the membrane surface and drive the fixture to perform large-range pose adjustments, thereby efficiently eliminating macroscopic warping and wrinkles. In the fine adjustment step, the system obtains the precise shape through stereo vision 3D reconstruction, and calculates the fine adjustment amount of each fixture by combining a pre-calibrated sensitivity matrix and a regularization optimization algorithm, achieving micron-level correction of local defects. In the closed-loop control step, the system dynamically adjusts the control weights of coarse and fine adjustments according to the iteration process, realizing a smooth transition and continuous optimization of the control strategy, and finally forming a stable closed loop that can cope with dynamic working conditions, significantly improving the flatness accuracy, response speed, and working condition adaptability of the coating process.

[0052] Before performing the coarse adjustment step, the following calibration and self-test steps must be completed in the method of this invention:

[0053] 1. Camera-projection calibration: Using a checkerboard or standard calibration board, complete the camera's intrinsic parameters (…). ), extrinsic parameters (rotation matrix) The phase-height calibration coefficients of the projection system are then determined using standard step blocks of known height. (The translation vector is used for calibration, followed by epipolar and distortion corrections.) Goodness of fit is ensured through linear fitting. The 95th percentile of the residual is less than 0.02 mm. The calibration coefficient is... The calculation formula is:

[0054] ;

[0055] in, Parameters that minimize the objective function The possible values ​​of ; For all data samples Summation; For the first The known reference height difference for each sample; For the first Phase difference obtained from individual sample measurements; These are the candidate coefficients to be tried during the optimization process; Indicates the first The squared residuals of each sample.

[0056] 2. Fixture-topography sensitivity matrix Identify:

[0057] For the A small disturbance is applied to the clamp. Simultaneously, the change in membrane height was collected. The sensitivity matrix is ​​constructed using the following formula. :

[0058] ;

[0059] If matrix rank Less than the number of fixtures It is necessary to increase the disturbance amplitude or optimize the fixture layout.

[0060] 3. Power-on self-test:

[0061] The system of this invention completes a self-test process within 30 seconds of startup, including: lens cleanliness detection (based on image contrast judgment), exposure and saturation check, and calibration parameter drift detection (required in this test). The self-test checks include: deviation from historical versions less than 2%, fixture origin homing, and motion boundary checks. Failure of any of these self-tests will trigger an alarm and prevent entry into the work cycle.

[0062] The coarse adjustment step involves monitoring and adjusting interference fringes globally, including the following steps:

[0063] 1. Image Acquisition and Processing: The projection device projects grating stripes onto the membrane surface at an incident angle of approximately 75°. The first camera above acquires the stripe image. After bandpass filtering enhancement, phase extraction and unpacking are performed using Hilbert transform or Fourier transform methods to obtain a continuous phase distribution. .

[0064] 2. Height Conversion and Index Calculation: Using Calibration Coefficients Convert the phase difference into a height increment: Thus, the membrane height distribution is obtained. Calculate the following global flatness indices:

[0065] Root mean square error: ;

[0066] in, The measurement area; For pixel coordinates The height value measured at the location; For the entire measurement area All height values The arithmetic mean, i.e. , Indicates the measurement area The total number of pixels within the region, i.e., the area of ​​the region; Root mean square error (RMSE) represents the deviation of the height of a single pixel from the average height. Flatness error is used to measure the overall deviation or fluctuation of the membrane surface height distribution relative to the average height. A smaller value indicates a flatter surface with fewer prominent highs or lows.

[0067] Flatness error (peak-valley value): ;

[0068] in, Indicates the measurement area The maximum value among all height values ​​within; Indicates the measurement area The minimum of all height values ​​within the area. Flatness error describes the maximum difference in elevation between the highest and lowest points within the entire measurement area. It reflects the total range of surface undulations.

[0069] Maximum gradient: ;

[0070] in, The gradient operator for the height function, The maximum gradient measures the inclination or slope of the steepest part of the surface. An excessively large value indicates a sharp change in surface height, which may lead to stress concentration or poor adhesion.

[0071] 3. Global PID control: As the main control input, a PID controller is used to output the global pose correction. :

[0072] ;

[0073] in, , , These are proportional gain, integral gain, and derivative gain, respectively. To control deviation; This is the cumulative sum of the deviations; The duration of the cycle; This is the deviation value from the previous control cycle; this embodiment uses... , , The controller must have anti-integral saturation and output limiting functions.

[0074] 4. Stage transition criterion: When and When the system continues to decrease for three consecutive control cycles, it determines that the coarse adjustment is complete and proceeds to the fine adjustment step.

[0075] The fine-tuning process involves 3D reconstruction monitoring and local fine-tuning, including the following steps:

[0076] 1. Three-dimensional topography reconstruction: Six stereo cameras simultaneously acquire images from both left and right eyes. After epipolar correction, the disparity map is calculated. Depth information can then be obtained through the following formula. and relative to the target morphology Height distribution :

[0077] ;

[0078] in, For the camera's focal length, Used as the baseline.

[0079] 2. Construct the weight matrix: Define the weight matrix Taking all factors into consideration:

[0080] Geometric weights: based on local curvature Normalized weighting results in higher weights for regions with high curvature.

[0081] Process weighting: Assigning higher weights to key areas such as welds and functional zones (e.g., Other areas are .

[0082] Confidence weights: Reduce the weight of occluded or low-texture areas in the image.

[0083] 3. Optimize the solution by adjusting the fine-tuning amount: adjust the current morphology With the target morphology The residuals were obtained by comparison. The optimal fine-tuning amount of the fixture is calculated by solving the following regularized least squares problem. :

[0084] ;

[0085] Its analytical solution is:

[0086] ;

[0087] in, This is the fixture-topography sensitivity matrix; For smoothing operators; Represents the L2 norm; and This is the regularization coefficient; the default value for the regularization parameter is [value to be filled in]. , The adaptive range is recommended to be , .

[0088] 4. Convergence Criterion: The system determines that the fine-tuning has converged and enters the hold state when the following conditions are met simultaneously:

[0089] ;

[0090] ;

[0091] ;

[0092] If there is no significant improvement after 5 consecutive iterations, the system reverts to the coarse-tuning step and reconstructs the weight matrix and region of interest.

[0093] In the closed-loop control step, closed-loop control and dynamic scheduling are performed, including the following steps:

[0094] 1. Weighted Scheduling: A continuous weighting function is used to smoothly transition between coarse and fine control values.

[0095] , ;

[0096] in, For coarse weighting; is the fine-tuning weight; is the number of iterations; κ is the decay constant.

[0097] To facilitate engineering implementation, discrete proportional scheduling is adopted:

[0098] Number of iterations Coarse adjustment: Fine adjustment weight ratio = ;

[0099] Number of iterations Coarse adjustment: Fine adjustment weight ratio = ;

[0100] Number of iterations Coarse adjustment: Fine adjustment weight ratio = .

[0101] If an indicator rebound is detected (e.g.) Or thermal disturbances, the weighting ratio is temporarily adjusted back to Until the system returns to stability.

[0102] 2. State Machine Management: The system operates on a state machine with the following process: INIT (initialization) → COARSE (coarse tuning) → FINE (fine tuning) → HOLD (hold). Any exception (such as occlusion, oversaturation, or out-of-bounds) will trigger a state transition to SAFE_PAUSE (safe pause), and after reset, it will return to the COARSE state.

[0103] Furthermore, the method of the present invention also includes a safety constraint and dynamic compensation step, specifically including:

[0104] 1. Execution Constraints: Limit the amplitude and rate of global adjustment commands and fine-tuning values.

[0105] Single-step adjustment amount meets ;

[0106] Adjustment rate satisfies , The adjustment rate of the fixture;

[0107] Trip constraints: , Rotation angle .

[0108] 2. Safety Protection: Soft limit switches, geometric collision cone detection, and emergency stop functions are enabled. Any trigger signal will... Internal response. Simultaneously, it achieves photothermal safety interlocking with the laser / preheating system, prohibiting laser illumination when the vision system is not in HOLD state.

[0109] 3. Dynamic feedforward compensation:

[0110] Motion perturbation feedforward: Real-time subscription of end effector velocity from robot controller and acceleration And generate feedforward compensation amount This is to counteract the disturbances caused by the movement of the embodied robot.

[0111] Thermally induced deformation prediction: Predicting the deformation at the next time step using online recursive least squares method or lightweight deep neural network. and use it as compensation ( Add it to the optimization objective. This is the compensation coefficient.

[0112] Vibration resistance processing: Apply bandpass filtering to the acquired height data to remove vibration. High-frequency vibration noise; binomial smoothing filter is used for executing commands: .

[0113] Example 2: This invention provides a dynamic monitoring and optimization system for the flatness of curved surface coating on a unibody robot. This system is used to implement the dynamic monitoring and optimization method for the flatness of curved surface coating on a unibody robot described in Example 1 above, specifically including:

[0114] The optical monitoring subsystem includes a projection device for projecting interference fringes, a first camera for acquiring fringe images, and a stereo vision system for three-dimensional topography reconstruction.

[0115] The multi-degree-of-freedom fixture subsystem includes an electric three-dimensional assembly stage and multiple independently controlled fixtures mounted thereon for fixing and adjusting the pose of the flexible membrane material.

[0116] The control subsystem, which communicates with the optical monitoring subsystem and the multi-degree-of-freedom fixture subsystem, is configured to perform coarse adjustment steps, fine adjustment steps, and closed-loop control steps.

[0117] like Figures 2 to 4 As shown, the hardware system upon which the implementation of this invention depends mainly consists of the following components:

[0118] 1. Fixed base plate, 2. Horizontal lead screw moving platform, 3. Horizontal slide table, 4. Base plate, 5. Stereo camera fixture, 6. Stereo camera, 7. Support device, 8. First servo motor, 9. Electric three-dimensional combination table, 10. Fixture, 11. Flexible electronic product (flexible film material), 12. Processed parts, 13. Second servo motor, 14. Vertical slide table, 15. Laser profilometer, 16. Vertical lead screw moving platform, 17. Projection device, 18. Tripod.

[0119] It should be noted that the first camera and the stereo vision system mentioned in the method of this invention are implemented by the same high-resolution stereo camera 6 in this embodiment. This camera serves as the first camera in the coarse adjustment stage to acquire interference fringe images, and as the stereo vision system in the fine adjustment stage to acquire multi-angle images for 3D reconstruction. This integrated design reduces system complexity and cost. The function of the first controller is implemented by the global PID control algorithm module in the control system of this embodiment.

[0120] The functions and layout relationships of each core component are as follows:

[0121] Fixed base plate 1: Serves as the supporting foundation for the entire system, ensuring the stability of the overall structure.

[0122] Multi-DOF Fixture Subsystem: The core adjustment unit consists of an electrically operated three-dimensional assembly stage 9 and multiple fixtures 10 distributed on it. Each fixture has six degrees of freedom (6DoF) fine-tuning capability (translation). , , With corner , , The pose is denoted as It can be independently controlled to achieve multi-point dynamic fixation and differentiated tension application of flexible membrane materials.

[0123] Optical monitoring subsystem:

[0124] Projection device 17: Fixed above the flexible membrane material at an angle of incidence. It is used to project grating interference fringes onto the film surface.

[0125] Stereo camera 6: Mounted directly above the flexible membrane material and secured by stereo camera clamp 5. Its baseline... The preferred focal length is 100-150mm, and the preferred working distance is 300-600mm. This camera can acquire interference fringe images for coarse adjustment, and can also acquire left and right eye images for fine adjustment of 3D reconstruction.

[0126] The vertical lead screw moving platform 16 and the horizontal lead screw moving platform 2 are driven by the first servo motor 8 and the second servo motor 13, and are used to drive the stereo camera 6 or the fixture 10 for precise positioning.

[0127] The control subsystem (not directly shown in the figure) includes an image acquisition card, a motion control card, and a host computer. It is responsible for receiving camera data and running monitoring and control algorithms. Its internal global PID control algorithm module corresponds to the first controller in this invention, used for global adjustments during the coarse-tuning stage.

[0128] This embodiment provides a dynamic monitoring and optimization system for the flatness of curved surface coating on a unibody robot, used to implement the aforementioned method for dynamic monitoring and optimization of the flatness of curved surface coating on a unibody robot. Therefore, the specific implementation of the system for dynamic monitoring and optimization of the flatness of curved surface coating on a unibody robot can be found in the previous section on the implementation of the method for dynamic monitoring and optimization of the flatness of curved surface coating on a unibody robot. To avoid redundancy, it will not be repeated here.

[0129] Example 3: This embodiment of the invention provides a computer storage medium storing a computer software product. The computer software product includes several instructions to cause a computer device to execute the above-described method for dynamic monitoring and optimization of the flatness of the embodied robot curved surface coating.

[0130] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0131] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0132] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0133] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A kind of embodied robot curved surface film covering flatness dynamic monitoring and optimization method, it is characterized in that, The method comprises the following steps: A coarse adjustment step: projecting an interference fringe onto the surface of the flexible film material fixed on the clamp by a projection device, and collecting a fringe image by a first camera; calculating a first height distribution of the film surface based on the phase change of the fringe image, and calculating a global flatness index based on the first height distribution; generating a global adjustment instruction by a first controller based on the global flatness index to drive the clamp to perform pose adjustment, so that the global flatness index meets a first threshold; A fine adjustment step: obtaining a three-dimensional morphology of the surface of the flexible film material by a stereo vision system to obtain a second height distribution; obtaining a residual error by comparing the second height distribution with a target morphology, and calculating a fine adjustment amount of each clamp based on the residual error and a pre-calibrated clamp-morphology sensitivity matrix by an optimization model with regularization to generate a second control instruction to drive the clamp to perform the fine adjustment amount, so that a flatness index corresponding to the second height distribution meets a second threshold; The closed-loop control step: during the whole coating process, the coarse adjustment step and the fine adjustment step are executed cyclically, and the control weights of the two steps are dynamically adjusted according to the iteration process, and the weights are switched between the coarse adjustment step and the fine adjustment step through a weight scheduling function, the weight scheduling function is: coarse adjustment weight , fine adjustment weight , wherein is the iteration number, and κ is the attenuation constant; the discrete execution ratio is: when the iteration number , the coarse adjustment weight: fine adjustment weight = 0.80:0.20; , the coarse adjustment weight: fine adjustment weight is 0.50:0.50; , the coarse adjustment weight: fine adjustment weight is 0.10:0.90, realizing the smooth transition from coarse adjustment to fine adjustment and dynamic closed-loop control.

2. The embodied robot curved surface coating flatness dynamic monitoring and optimization method according to claim 1, characterized in that, In the coarse adjustment step, the first height distribution of the film surface is calculated based on the phase change of the fringe image, and the global flatness index is calculated based on the first height distribution, which specifically comprises: The phase distribution is obtained from the stripe image using a phase extraction and unpacking algorithm. And using calibration coefficients Through formula The change in height distribution was calculated. The global flatness index includes the root mean square error of the height distribution. Flatness error and maximum gradient ; Wherein ； ； ； wherein is the measurement area; is the height value measured at the pixel coordinate ; is the arithmetic mean of all height values within the entire measurement area , i.e. , denotes the total number of pixel points within the measurement area , i.e. the area of the area; denotes the deviation of the height of a single pixel point from the average height; denotes the maximum value of all height values within the measurement area ; denotes the minimum value of all height values within the measurement area ; is the gradient operator of the height function, .

3. The embodied robotic method of dynamic monitoring and optimization of the flatness of a curved film according to claim 2, wherein, In the coarse adjustment step, when the less than or equal to a first threshold value and when the number of times of decrease is continuous, it is determined to enter the fine adjustment step.

4. The embodied robot curved surface coating flatness dynamic monitoring and optimization method according to claim 1, characterized in that, In the coarse adjustment step, the first controller is a PID controller, and the global adjustment instruction generated by the first controller comprises a global pose correction amount The calculation formula of the global pose correction amount is as follows: ； wherein, , , are a proportional gain, an integral gain and a differential gain, respectively; is a control deviation; is a cumulative sum of the deviation; is a cycle length; is a deviation value of a previous control cycle.

5. The embodied robot curved surface coating flatness dynamic monitoring and optimization method according to claim 1, characterized in that, In the fine adjustment step, the mathematical expression of the optimization model is: ； wherein, is the fine tuning amount of the clamp; is the residual; is the clamp-topography sensitivity matrix; is the weight matrix; and is the regularization coefficient; is the smoothing operator; denotes the L2 norm; its analytical solution is: ； wherein denotes the transpose.

6. The embodied robotic method of dynamic monitoring and optimization of the flatness of a curved film according to claim 5, wherein, In the fine adjustment step, when the root mean square error is less than a second threshold and , and the single-step adjustment amount of the jig , it is determined that the fine adjustment step converges, wherein is a preset positive threshold.

7. The embodied robot method of dynamic monitoring and optimization of the flatness of a curved surface coated with a film according to claim 1, characterized in that, The method further comprises performing a constraint step to limit the amplitude and the rate of the global adjustment instruction and the fine adjustment amount, wherein the single-step adjustment amount satisfies , and the adjustment rate satisfies ; and limiting the stroke and the rotation angle of the clamp, wherein is the fine adjustment amount of the clamp, is the adjustment rate of the clamp.

8. A dynamic monitoring and optimization system for the flatness of a curved film of a robot with a body, characterized in that, The system is used to implement the embodied robot curved surface film covering flatness dynamic monitoring and optimization method of any one of claims 1 to 7, comprising: An optical monitoring subsystem comprising a projection device for projecting an interference fringe, a first camera for collecting a fringe image, and a stereo vision system for three-dimensional morphology reconstruction; A multi-degree-of-freedom clamp subsystem comprising an electric three-dimensional combined table and a plurality of independently controlled clamps mounted thereon for fixing and adjusting the pose of the flexible film material; A control subsystem in communication connection with the optical monitoring subsystem and the multi-degree-of-freedom clamp subsystem, configured to perform the coarse adjustment step, the fine adjustment step and the closed-loop control step.

9. A computer storage medium, characterized in that, The computer storage medium stores a computer software product, and the computer software product comprises a plurality of instructions for causing a computer device to execute the embodied robot curved surface film covering flatness dynamic monitoring and optimization method of any one of claims 1 to 7.

Images