Motion track smooth optimization method and device, equipment and storage medium
By collecting and optimizing two-dimensional discrete point data, and using regression models and Hermite spline curves to optimize the motion trajectory, the stability and consistency issues of flat-top beams in large-format glass plate processing were solved, enabling efficient industrial production.
Patent Information
- Application Number
- CN202511514424.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-22
- Publication Date
- 2026-01-27
AI Technical Summary
In the industrial production of perovskite solar cells, the shallow depth of focus of flat-top beams is difficult to control when processing large-format glass panels, leading to problems with processing stability and consistency. Furthermore, high-frequency command control causes mechanical resonance in the slave axis servo system, affecting processing quality.
By collecting distance data from the Z-axis to the glass surface, and combining it with the X-axis scribing distance and sampling density, two-dimensional discrete point data is established. The motion trajectory is optimized using regression models and Hermite spline curves to suppress oscillations of the slave axis during high-speed operation of the master axis and maintain the laser focus on the optimal processing plane on the glass surface.
It improves the stability and reliability of the flat-top light, suppresses the oscillation of the slave axis, enhances the overlap between the actual trajectory and the theoretical trajectory, and realizes the industrial processing of high-efficiency large-format glass plates.
Smart Images
Figure CN121411244A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of laser cutting and etching technology, and in particular to a method, apparatus, device and storage medium for smoothing and optimizing motion trajectory. Background Technology
[0002] In the laser dicing and laser scribing processes of industrial production of perovskite solar cells, Gaussian beams, due to their energy distribution characteristics of high energy at the center and low energy at the edges, are prone to problems such as increased edge roughness, significant burrs, prominent crater effect, and excessive concentration of heat-affected zones, severely restricting processing accuracy and cell performance. In contrast, flat-top beams have a more uniform energy distribution than Gaussian beams, thus offering advantages in material processing such as better edge effects and cleaner material removal. They can significantly reduce crater and burr effects, significantly improve edge morphology, suppress material sputtering, and reduce thermal damage, thereby exhibiting superior processing quality. However, they also have the disadvantage of shallow depth of focus. When processing large-format glass sheets, it is extremely difficult to control the surface flatness of the product within the depth of focus range of flat-top beams (usually less than 0.1 mm). Therefore, the application of flat-top beams is mainly for experimental use on small-format equipment in laboratories and is difficult to directly apply to the industrial manufacturing of large-format glass substrates on production lines.
[0003] Therefore, to address this shortcoming of flat-top beam applications in large-format laser segmentation and scribing of perovskites, there exists an automatic focusing motion trajectory calculation method that uses a laser rangefinder to measure the distance between the laser head and the glass plate in real time and adjusts the Z-axis position in real time to ensure that the laser intersection point is always on the glass plate. The basic principle is to use the height fluctuation data of the glass plate surface on the X-axis motion area scanned by the laser rangefinder to pre-plan the Z-axis motion trajectory based on the data. Correspondingly, when the main axis X-axis moves, the secondary axis Z-axis follows the main axis, ensuring that the focus of the flat-top beam is always on the glass surface.
[0004] In data acquisition systems based on encoder-based fixed-pulse triggering mechanisms, the reconstruction accuracy of discrete sampled data directly determines the realism of the glass surface morphology fitting. Theoretically, sampling density is positively correlated with reconstruction accuracy; a higher sampling rate can more accurately approximate the continuous features of surface undulations.
[0005] However, large-format processing for industrial applications typically requires a processing cycle of less than 30 seconds for the entire glass sheet, and a laser speed of at least 1.5 m / s. Under these high-speed processing conditions, high-density sampling of the glass surface morphology is necessary to achieve high-precision focus tracking.
[0006] However, the denser the sampling points, the more likely the high-frequency commands in the motion control system will frequently cause the slave axis to perform forward and reverse compensation movements at high speeds, which will easily trigger mechanical resonance in the slave axis servo system, introduce nonlinear vibration errors, and cause significant oscillations in the Z-axis during the following motion. This will negatively affect the stability of the flat-top light processing process and the surface forming quality, seriously threatening the processing stability and consistency. Summary of the Invention
[0007] The purpose of this invention is to provide a motion trajectory smoothing optimization method, apparatus, device, and storage medium. By smoothing and optimizing the motion curve of the slave axis, the following oscillation caused by high-frequency commands during the high-speed operation of the master axis is suppressed. This effectively improves the overlap between the actual trajectory and the theoretical trajectory of the slave axis, which is beneficial to keeping the laser focus on the optimal processing plane of the glass surface and improving the stability and reliability of the flat-top beam.
[0008] To achieve the above objectives, this invention discloses a motion trajectory smoothing optimization method, which includes: The X-axis is controlled to drive the Z-axis to translate at a scribing speed V, and the rangefinder is controlled to collect distance data from the Z-axis to the glass surface. Obtain the line distance data of the X-axis, and combine the collected distance data and sampling density ρ to obtain m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis; Calculate all trajectory angles θ and the normalization parameter k for each angle in the motion trajectory obtained from two-dimensional discrete point data; The two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k are combined to obtain multiple feature vectors in units of angle. Multiple feature vectors are input into the regression model, and multiple recommended parameters β corresponding to the trajectory angle θ are obtained from the regression model output; Based on multiple recommended parameters β and the normalized parameter k for each included angle, multiple tension control parameters α corresponding to the included angle θ of the trajectory are calculated; The tangential vector of each two-dimensional discrete point is calculated using multiple tension control parameters α and two-dimensional discrete point data. Multiple Hermite spline curves are constructed using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point to obtain the trajectory optimization curve.
[0009] Furthermore, the regression model is a gradient boosting tree model.
[0010] Furthermore, after establishing multiple Hermite spline curves using two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to obtain the trajectory optimization curve, the process further includes: The X-axis is controlled to drive the Z-axis to translate, while the Z-axis is controlled to move according to the trajectory optimization curve, and the movement position data of the X-axis and Z-axis are collected in real time. The root mean square error is calculated based on the actual collected X-axis and Z-axis movement position data and m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis, and the calculated root mean square error is used as the loss function of the regression model. Optimize the parameters of the regression model in the direction that reduces the loss function value.
[0011] Furthermore, the calculation of all trajectory angles θ in the motion trajectory obtained from two-dimensional discrete point data, and the normalization parameter k for each angle, includes: The trajectory angle θ is calculated using the first formula, which is:
[0012] in, , , It is two-dimensional discrete point data; The normalized parameter k of the included angle is calculated using the second formula, which is:
[0013] in, The angle θ between the trajectory and the calculated normalized parameter k is... It is the minimum value of the included angle θ among all trajectories. It is the maximum value of the included angle θ among all trajectories.
[0014] Furthermore, the calculation of multiple tension control parameters α corresponding to the trajectory angle θ based on multiple recommended parameters β and the normalized parameter k of each angle includes: The tension control parameter α is calculated using the third formula, which is:
[0015] in, is the normalized parameter for the included angle, and β is the recommended parameter and is greater than zero.
[0016] Furthermore, the calculation of the tangential vector for each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data includes: Calculate the tangent vector of the first two-dimensional discrete point using the fourth formula. The fourth formula is:
[0017] in, For the data of the first two-dimensional discrete point, This is the data for the second two-dimensional discrete point; Calculate the tangent vector of the last two-dimensional discrete point using the fifth formula. The fifth formula is:
[0018] in, For the data of the last two-dimensional discrete point, This refers to the data from the second-to-last two-dimensional discrete point; Calculate the tangent vector between the first and last two-dimensional discrete points using the sixth formula. The sixth formula is:
[0019] in, , , It is two-dimensional discrete point data. and Representing vectors respectively The magnitude and vector The model, These are tension control parameters.
[0020] Furthermore, the step of establishing multiple Hermite spline curves using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point to obtain the trajectory optimization curve includes: Calculate each Hermite spline curve using the seventh formula. The seventh formula is: + +
[0021] Where n ranges from (1, m). It is two-dimensional discrete point data. The tangent vector of a two-dimensional discrete point. t is in the range of (0, 1), d is the distance from the starting point of the trajectory optimization curve to the current translation position of the Z-axis, and D is the distance from the starting point to the ending point of the trajectory optimization curve.
[0022] To achieve the above objectives, the present invention discloses a motion trajectory smoothing optimization device, comprising: The control module is used to control the X-axis to drive the Z-axis to translate at a scribing speed V, and to control the rangefinder to collect distance data from the Z-axis to the glass surface; The module is used to acquire the line distance data of the X-axis, and combined with the acquired distance data and sampling density ρ, m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis are obtained; The first calculation module is used to calculate all the trajectory angles θ and the normalization parameter k of each angle in the motion trajectory obtained from the two-dimensional discrete point data. The combination module is used to combine two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k to obtain multiple feature vectors in units of angle. The input module is used to input multiple feature vectors into the regression model and obtain multiple recommended parameters β output by the regression model corresponding to the trajectory angle θ. The second calculation module is used to calculate multiple tension control parameters α corresponding to the trajectory angle θ based on multiple recommended parameters β and the normalized parameter k of each angle. The third calculation module is used to calculate the tangential vector of each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data. A module is established to create multiple Hermite spline curves using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point, in order to obtain the trajectory optimization curve.
[0023] To achieve the above objectives, the present invention discloses an electronic device comprising: One or more processors; One or more memories are used to store one or more programs, which, when executed by the processor, enable the processor to implement the motion trajectory smoothing optimization method as described above.
[0024] To achieve the above objectives, the present invention discloses a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the motion trajectory smoothing optimization method as described above.
[0025] Compared with existing technologies, this invention first collects distance data from the Z-axis to the glass surface, and combines this with scribing distance data on the X-axis and sampling density ρ to obtain m sets of two-dimensional discrete point data. Then, multiple feature vectors established by combining two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, scribing speed V, trajectory angle θ, and normalization parameter k are input into a regression model to obtain recommended parameter β, and the tension control parameter α is calculated accordingly. Finally, the tangential vector of each two-dimensional discrete point is calculated, and multiple Hermite spline curves are established to obtain the trajectory optimization curve, thereby completing the smooth optimization of the slave axis motion curve. Based on the trajectory angle θ, each segment of the slave axis motion trajectory is optimized, which can suppress the following oscillation caused by high-frequency commands when the slave axis is running at high speed on the main axis, effectively improve the overlap between the actual trajectory and the theoretical trajectory of the slave axis, and help maintain the laser focus on the optimal processing plane on the glass surface, thereby improving the stability and reliability of the flat-top beam. Attached Figure Description
[0026] Figure 1 This is a flowchart of the motion trajectory smoothing optimization method according to an embodiment of the present invention.
[0027] Figure 2 This is a schematic diagram of the coordinates of two-dimensional discrete point data in the motion trajectory smoothing optimization method of this invention.
[0028] Figure 3 This is a schematic diagram of the motion trajectory along the Z-axis in the motion trajectory smoothing optimization method of this invention.
[0029] Figure 4 This is a block diagram of the motion trajectory smoothing optimization device according to an embodiment of the present invention.
[0030] Figure 5 This is a system diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation
[0031] To illustrate the technical content, structural features, objectives, and effects of the present invention in detail, the following description is provided in conjunction with the embodiments and accompanying drawings.
[0032] Example 1 Please see Figures 1 to 3 This invention discloses a method for smoothing and optimizing motion trajectories, comprising: 101. Control the X-axis to drive the Z-axis to translate at the scribing speed V, and control the rangefinder to collect the distance data from the Z-axis to the glass surface; 102. Obtain the line distance data of the X-axis, and combine the collected distance data and sampling density ρ to obtain m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis; It should be noted that in this embodiment, the distance data from the Z-axis to the glass surface, i.e., Zpos, is collected by triggering the rangefinder with an encoder. The frequency of the encoder triggering is the sampling density. The position where the X-axis drives the Z-axis to translate when the encoder triggers the rangefinder is obtained by acquiring the scribing distance data of the X-axis, i.e., Xpos. Furthermore, the obtained two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis are the two-dimensional discrete point positions that the Z-axis motion trajectory should ideally pass through. The closer the actual motion trajectory of the Z-axis is to this theoretical trajectory, the better the effect of maintaining the laser focus on the optimal processing plane of the glass surface, but it is not limited to this.
[0033] 103. Calculate all the trajectory angles θ and the normalization parameter k for each angle in the motion trajectory obtained from the two-dimensional discrete point data; Furthermore, the calculation of all trajectory angles θ in the motion trajectory obtained from the two-dimensional discrete point data, and the normalization parameter k for each angle, includes: 1031. Calculate the included angle θ of the trajectory using the first formula, which is:
[0034] in, , , It is two-dimensional discrete point data; 1032. Calculate the normalized parameter k of the included angle using the second formula, which is:
[0035] in, The angle θ between the trajectory and the calculated normalized parameter k is... It is the minimum value of the included angle θ among all trajectories. It is the maximum value of the included angle θ among all trajectories.
[0036] 104. Combine the two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k to obtain multiple feature vectors in units of angle. 105. Input multiple feature vectors into the regression model and obtain multiple recommended parameters β corresponding to the trajectory angle θ output by the regression model; Furthermore, the regression model is a gradient boosting tree model.
[0037] It is understandable that, in this embodiment, since there is no need to make complex linear assumptions about the features, the gradient boosting tree model can effectively capture complex nonlinear relationships and evaluate the importance of the features. The aforementioned process parameters and operating condition parameters are engineered to form a feature vector with the included angle as the unit. As an input parameter of the regression model, it is beneficial to adaptively adjust the predicted recommended β value based on the process parameters and operating conditions, so as to adaptively adjust the smoothness of the trajectory curve at different positions.
[0038] 106. Based on multiple recommended parameters β and the normalized parameter k for each included angle, multiple tension control parameters α corresponding to the included angle θ of the trajectory are calculated.
[0039] Furthermore, based on multiple recommended parameters β and the normalized parameter k for each included angle, multiple tension control parameters α corresponding to the trajectory included angle θ are calculated, including: 1061. Calculate the tension control parameter α using the third formula, which is:
[0040] in, is the normalized parameter for the included angle, and β is the recommended parameter and is greater than zero.
[0041] 107. Calculate the tangential vector of each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data; Furthermore, the tangential vector for each two-dimensional discrete point is calculated using multiple tension control parameters α and two-dimensional discrete point data, including: 1071. Calculate the tangent vector of the first two-dimensional discrete point using the fourth formula. The fourth formula is:
[0042] in, For the data of the first two-dimensional discrete point, This is the data for the second two-dimensional discrete point; 1072. Calculate the tangent vector of the last two-dimensional discrete point using the fifth formula. The fifth formula is:
[0043] in, For the data of the last two-dimensional discrete point, This refers to the data from the second-to-last two-dimensional discrete point; 1073. Calculate the tangent vector between the first and last two-dimensional discrete points using the sixth formula. The sixth formula is:
[0044] in, , , It is two-dimensional discrete point data. and Representing vectors respectively The magnitude and vector The model, These are tension control parameters.
[0045] 108. Use two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to establish multiple Hermite spline curves to obtain the trajectory optimization curve.
[0046] Furthermore, multiple Hermite spline curves are constructed using two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to obtain the trajectory optimization curve, including: 1081. Calculate each Hermite spline curve using the seventh formula. The seventh formula is: + +
[0047] Where n ranges from (1, m). It is two-dimensional discrete point data. Let be the tangent vector of a two-dimensional discrete point, and t be the normalized arc length parameter. t is in the range of (0, 1), d is the distance from the starting point of the trajectory optimization curve to the current translation position of the Z-axis, and D is the distance from the starting point to the ending point of the trajectory optimization curve.
[0048] Understandably, Hermite spline curves are introduced to plan the motion trajectory along the Z-axis. Hermite splines achieve smoothness by controlling the derivatives of the vertices. Therefore, the tension control parameter α is used to adjust the magnitude of the tangential vector at each vertex. The larger α is, the smaller the derivative, and the smoother the curve. Furthermore, the smoothness of the curve is adaptively adjusted at various points along the motion trajectory based on the included angles. For example, ... Figure 3 The large included angle θ1 shown reduces the smoothness of the motion trajectory at this point, ensuring a smooth motor rotation on the Z-axis and a natural transition in the motion trajectory. Figure 3 The small included angle θ2 shown enhances the smoothness of the motion trajectory at this point, preventing abrupt changes in motion. Specifically, by adaptively adjusting parameter β, the values of each tension control parameter α are optimized, dynamically adjusting the smoothness of the Z-axis motion trajectory. When the scribing speed on the X-axis needs to reach 1.5 m / s, the oscillation phenomenon caused by tracking high-frequency, abrupt position commands on the Z-axis is effectively suppressed. At the same time, the overlap between the actual trajectory and the theoretical trajectory of the Z-axis is effectively improved, ensuring that the motion trajectory passes through all vertices (two-dimensional discrete point data of the X-axis and Z-axis). This facilitates the dynamic maintenance of the flat-top light focus on the optimal processing plane on the glass surface, realizing the stable and reliable application of flat-top light technology in high-efficiency, large-format industrial production.
[0049] Furthermore, after constructing multiple Hermite spline curves using two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to obtain the trajectory optimization curve, the process also includes: 109. Control the X-axis to drive the Z-axis translation, and simultaneously control the Z-axis movement according to the trajectory optimization curve, and collect the movement position data of the X-axis and Z-axis in real time; 110. Calculate the root mean square error based on the actual collected X-axis and Z-axis movement position data and m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis, and use the calculated root mean square error as the loss function of the regression model. Furthermore, the root mean square error is calculated using the eighth formula and used as the loss function for the regression model. The eighth formula is:
[0050] Where m is the number of points in the two-dimensional discrete data, Z is the actual Z-axis movement position data collected, and Z' is Zpos in the two-dimensional discrete data.
[0051] 111. Optimize the parameters of the regression model in the direction that reduces the loss function value.
[0052] Understandably, in this embodiment, the movement of the Z-axis during actual scribing is controlled based on the obtained trajectory optimization curve. The encoder triggers the acquisition of the actual movement position data of the X-axis and Z-axis during processing using the same sampling density as when acquiring two-dimensional discrete point data. This data is then used to calculate the loss function value of the regression model. Based on existing methods for optimizing the regression model in the direction of decreasing the loss function value, the weights and other parameters of the decision tree in the gradient boosting tree model are adjusted to obtain new recommended parameters β and trajectory optimization curves in subsequent scribing processes, thereby achieving closed-loop optimization.
[0053] Compared with existing technologies, this invention first collects distance data from the Z-axis to the glass surface, and combines this with scribing distance data on the X-axis and sampling density ρ to obtain m sets of two-dimensional discrete point data. Then, multiple feature vectors established by combining two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, scribing speed V, trajectory angle θ, and normalization parameter k are input into a regression model to obtain recommended parameter β, and the tension control parameter α is calculated accordingly. Finally, the tangential vector of each two-dimensional discrete point is calculated, and multiple Hermite spline curves are established to obtain the trajectory optimization curve, thereby completing the smooth optimization of the slave axis motion curve. Based on the trajectory angle θ, each segment of the slave axis motion trajectory is optimized, which can suppress the following oscillation caused by high-frequency commands when the slave axis is running at high speed on the main axis, effectively improve the overlap between the actual trajectory and the theoretical trajectory of the slave axis, and help maintain the laser focus on the optimal processing plane on the glass surface, thereby improving the stability and reliability of the flat-top beam.
[0054] Example 2 Please see Figures 1 to 4 This invention discloses a motion trajectory smoothing optimization device, which includes: The control module 201 is used to control the X-axis to drive the Z-axis to translate at a scribing speed V, and to control the rangefinder to collect distance data from the Z-axis to the glass surface. Combined with module 202, it is used to obtain the line distance data of the X-axis, and combined with the collected distance data and sampling density ρ, it obtains m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis; The first calculation module 203 is used to calculate all the trajectory angles θ and the normalization parameter k of each angle in the motion trajectory obtained based on the two-dimensional discrete point data. The combination module 204 is used to combine two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k to obtain multiple feature vectors in terms of angle. The input module 205 is used to input multiple feature vectors into the regression model and obtain multiple recommended parameters β output by the regression model corresponding to the trajectory angle θ. The second calculation module 206 is used to calculate multiple tension control parameters α corresponding to the trajectory angle θ based on multiple recommended parameters β and the normalized parameter k of each angle. The third calculation module 207 is used to calculate the tangential vector of each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data; Module 208 is used to establish multiple Hermite spline curves using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point to obtain the trajectory optimization curve.
[0055] Example 3 Please see Figures 1 to 5 This invention discloses an electronic device comprising: One or more processors 301; One or more memories 302 are used to store one or more programs, which, when executed by a processor, enable the processor to implement the motion trajectory smoothing optimization method as described above.
[0056] Example 4 This application discloses a computer-readable storage medium storing a program thereon, which, when executed by a processor, implements the motion trajectory smoothing optimization method as described above.
[0057] Example 5 This application discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned motion trajectory smoothing optimization method.
[0058] It should be understood that, in the embodiments of this application, the processor may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0059] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by hardware related to computer program instructions. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.
[0060] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.
Claims
1. A method for smoothing and optimizing motion trajectories, characterized in that, include: The X-axis is controlled to drive the Z-axis to translate at a scribing speed V, and the rangefinder is controlled to collect distance data from the Z-axis to the glass surface. Obtain the line distance data of the X-axis, and combine the collected distance data and sampling density ρ to obtain m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis; Calculate all trajectory angles θ and the normalization parameter k for each angle in the motion trajectory obtained from two-dimensional discrete point data; The two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k are combined to obtain multiple feature vectors in units of angle. Multiple feature vectors are input into the regression model, and multiple recommended parameters β corresponding to the trajectory angle θ are obtained from the regression model output; Based on multiple recommended parameters β and the normalized parameter k for each included angle, multiple tension control parameters α corresponding to the included angle θ of the trajectory are calculated; The tangential vector of each two-dimensional discrete point is calculated using multiple tension control parameters α and two-dimensional discrete point data. Multiple Hermite spline curves are constructed using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point to obtain the trajectory optimization curve.
2. The motion trajectory smoothing optimization method according to claim 1, characterized in that, The regression model is a gradient boosting tree model.
3. The motion trajectory smoothing optimization method according to claim 1, characterized in that, After establishing multiple Hermite spline curves using two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to obtain the trajectory optimization curve, the process further includes: The X-axis is controlled to drive the Z-axis to translate, while the Z-axis is controlled to move according to the trajectory optimization curve, and the movement position data of the X-axis and Z-axis are collected in real time. The root mean square error is calculated based on the actual collected X-axis and Z-axis movement position data and m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis, and the calculated root mean square error is used as the loss function of the regression model. Optimize the parameters of the regression model in the direction that reduces the loss function value.
4. The motion trajectory smoothing optimization method according to claim 1, characterized in that, The calculation, based on two-dimensional discrete point data, includes all trajectory angles θ and the normalized parameter k for each angle in the motion trajectory. The trajectory angle θ is calculated using the first formula, which is: in, , , It is two-dimensional discrete point data; The normalized parameter k of the included angle is calculated using the second formula, which is: in, The angle θ between the trajectory and the calculated normalized parameter k is... It is the minimum value of the included angle θ among all trajectories. It is the maximum value of the included angle θ among all trajectories.
5. The motion trajectory smoothing optimization method according to claim 1, characterized in that, The multiple tension control parameters α corresponding to the trajectory angle θ are calculated based on multiple recommended parameters β and the normalized parameter k of each angle, including: The tension control parameter α is calculated using the third formula, which is: in, is the normalized parameter for the included angle, and β is the recommended parameter and is greater than zero.
6. The motion trajectory smoothing optimization method according to claim 1, characterized in that, The calculation of the tangential vector for each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data includes: Calculate the tangent vector of the first two-dimensional discrete point using the fourth formula. The fourth formula is: in, For the data of the first two-dimensional discrete point, This is the data for the second two-dimensional discrete point; Calculate the tangent vector of the last two-dimensional discrete point using the fifth formula. The fifth formula is: in, For the data of the last two-dimensional discrete point, This refers to the data from the second-to-last two-dimensional discrete point; Calculate the tangent vector between the first and last two-dimensional discrete points using the sixth formula. The sixth formula is: in, , , It is two-dimensional discrete point data. and Representing vectors respectively modulus and vector The model, These are tension control parameters.
7. The motion trajectory smoothing optimization method according to claim 1, characterized in that, The process of establishing multiple Hermite spline curves using two-dimensional discrete point data and the tangent vector of each two-dimensional discrete point to obtain the trajectory optimization curve includes: Calculate each Hermite spline curve using the seventh formula. The seventh formula is: + + Where n ranges from (1, m). It is two-dimensional discrete point data. The tangent vector of a two-dimensional discrete point. t is in the range of (0, 1), d is the distance from the starting point of the trajectory optimization curve to the current translation position of the Z-axis, and D is the distance from the starting point to the ending point of the trajectory optimization curve.
8. A motion trajectory smoothing optimization device, characterized in that, include: The control module is used to control the X-axis to drive the Z-axis to translate at a scribing speed V, and to control the rangefinder to collect distance data from the Z-axis to the glass surface; The module is used to acquire the line distance data of the X-axis, and combined with the acquired distance data and sampling density ρ, m sets of two-dimensional discrete point data (Xpos, Zpos) of the X-axis and Z-axis are obtained; The first calculation module is used to calculate all the trajectory angles θ and the normalization parameter k of each angle in the motion trajectory obtained from the two-dimensional discrete point data. The combination module is used to combine two-dimensional discrete point data (Xpos, Zpos), sampling density ρ, line drawing speed V, trajectory angle θ, and normalization parameter k to obtain multiple feature vectors in units of angle. The input module is used to input multiple feature vectors into the regression model and obtain multiple recommended parameters β output by the regression model corresponding to the trajectory angle θ. The second calculation module is used to calculate multiple tension control parameters α corresponding to the trajectory angle θ based on multiple recommended parameters β and the normalized parameter k of each angle. The third calculation module is used to calculate the tangential vector of each two-dimensional discrete point using multiple tension control parameters α and two-dimensional discrete point data. A module is established to create multiple Hermite spline curves using two-dimensional discrete point data and the tangential vector of each two-dimensional discrete point, in order to obtain the trajectory optimization curve.
9. An electronic device, characterized in that, include: One or more processors; One or more memories are used to store one or more programs, which, when executed by the processor, cause the processor to implement the motion trajectory smoothing optimization method as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the motion trajectory smoothing optimization method as described in any one of claims 1 to 7.