A laser multi-axis fly machining path planning method of adaptive weighted filtering

Abstract

The present application relates to the technical field of intelligent control, and more particularly to a laser multi-axis flight machining path planning method based on adaptive weighted filtering, which comprises inputting a three-dimensional curved surface model and a target two-dimensional path, setting constraint thresholds of defocusing amount, laser incidence angle and galvanometer scanning range, setting a lower bound of a filtering factor, generating a midpoint filtering factor and a weight matrix of an initialized filtering factor, performing weighted filtering on the target two-dimensional path through an adaptive weighted filtering target function to obtain a smoothed two-dimensional path, projecting the smoothed two-dimensional path onto the three-dimensional curved surface to obtain a platform machining path, calculating constraint parameters and constraint margins, generating segment filtering factor weights and updating the weight matrix, repeating the steps until iteration is completed, judging whether the platform machining path meets the constraints, updating a search interval, repeating the steps to output a final platform machining path, and calculating a galvanometer machining path. The method improves the path smoothing degree, thereby improving the average feeding speed of the motion platform and the machining efficiency.

Description

An Adaptive Weighted Filtering Method for Laser Multi-Axis Flight Processing Path Planning Technical Field

[0001] This invention relates to the field of intelligent control technology, and in particular to an adaptive weighted filtering method for laser multi-axis flight processing path planning. Background Technology

[0002] During laser flight processing, the multi-axis motion platform and the galvanometer need to move in tandem to complete the surface texture processing of curved workpieces. The laser beam needs to maintain a stable spot shape on the workpiece surface. Therefore, the processing path planning needs to meet three major constraints: defocusing amount, laser incident angle, and galvanometer scanning range.

[0003] Existing technologies typically employ a single-step global filtering process to obtain a smoothed 2D path for the target. This filtered path is then projected onto a 3D surface to obtain the platform machining path. The defocusing amount, laser incident angle, and galvanometer scanning range are calculated to determine if constraints are met, outputting the target and platform machining paths that satisfy the constraints. Finally, the galvanometer machining path is calculated using the target path and the platform machining path. This method has significant drawbacks: First, the filtering factor in a single-step global filtering process can only select conservative values. A few constrained regions dominate the overall smoothness, while the smoothing potential of constrained regions cannot be fully explored, thus limiting further smoothing in these regions and restricting the improvement of platform feed speed. Second, it treats different path segments indiscriminately, lacking a constraint-margin-based filtering weight allocation mechanism. Indiscriminate smoothing of the entire path easily leads to over-smoothing in constrained regions, causing constraint overruns, or insufficient smoothing in constrained regions, resulting in low machining efficiency.

[0004] Therefore, there is an urgent need for a planning method that can adaptively adjust the smoothness of the path according to the constraints, so as to improve the efficiency and quality of multi-axis laser flying processing while satisfying all constraints. Summary of the Invention

[0005] The purpose of this invention is to propose an adaptive weighted filtering laser multi-axis flight machining path planning method. Under the premise that three types of constraints—defocus amount, incident angle, and galvanometer scanning range—are simultaneously satisfied, the method adaptively adjusts the segment filtering factor weights through constraint feedback to achieve adaptive allocation of path smoothness. This allows the smoothness to be automatically reduced in constrained areas to ensure that the path does not go out of bounds, while the smoothness is automatically increased in constrained areas. This shortens the platform machining path length as much as possible, reduces curvature and vibration impact, thereby improving the average feed speed and machining efficiency of the motion platform.

[0006] To achieve this objective, the present invention adopts the following technical solution:

[0007] An adaptive weighted filtering method for laser multi-axis flight processing path planning includes the following steps:

[0008] S1: Input the 3D surface model of the workpiece and the target 2D path in the workpiece coordinate system, and set the constraint thresholds for defocusing amount, laser incident angle, and galvanometer scanning range as follows: , , Set the lower bound of the filter factor λ. ;

[0009] S2: Determine the search interval for the filter factor that satisfies the constraints of defocusing amount, laser incident angle, and galvanometer scanning range. ;

[0010] S3: Based on the filter factor search interval The midpoint filter factor is generated using the logarithmic domain bisection method. Initialize the weight matrix of the filter factors;

[0011] S4: Based on midpoint filter factor The weight matrix of the initial filtering factor is used to perform weighted filtering on the target two-dimensional path F through the adaptive weighted filtering objective function to obtain the smooth two-dimensional path E;

[0012] Adaptive weighted filtering objective function:

[0013] ,

[0014] in, For the points on the target two-dimensional path, For points on a smooth 2D path, This indicates a deviation from the cost. Indicates a smooth penalty. represents the segment filter factor weight in the weight matrix W, and L represents the smoothing operator;

[0015] S5: Project the smooth 2D path onto the 3D surface to obtain the platform machining path, and calculate the constraint parameters and constraint margins. The constraint parameter is the defocus amount. laser incident angle and galvanometer scanning range Generate segment filter factor weights based on constraint margins Update the weight matrix; repeat steps S3 and S4 until the path converges or the maximum number of weight iterations is reached, and output the platform processing path;

[0016] S6: Determine whether the platform processing path output by S5 meets the constraints. If it does, then... Updated to Otherwise Updated to ;

[0017] S7: Repeat S3-S6 until the filter factor converges or the maximum number of iterations of the filter factor is reached, and output the final platform processing path;

[0018] S8: Calculate the galvanometer machining path based on the target machining path and the platform machining path. .

[0019] Furthermore, in step S4, For squared, robust loss functions, or absolute values; It is the squared or robust loss function, and L is the first-order difference, second-order difference, or spline second derivative.

[0020] Furthermore, in step S1:

[0021] Defocus amount is defined as the target processing path point Platform processing path points Unit normal vector at the platform processing path point Absolute distance in direction: defocus amount ;

[0022] The laser incident angle is defined as the normal vector of the target machining path point. platform processing path point normal vector The included angle, laser incident angle ;

[0023] The galvanometer scanning range is the maximum millimeter displacement of the U-axis and V-axis in the galvanometer focal plane coordinate system. , ;

[0024] The constraints of defocusing amount, laser incident angle, and galvanometer scanning range are simultaneously satisfied:

[0025] .

[0026] Furthermore, in step S2:

[0027] From the lower bound of the filter factor The process begins by multiplying candidate filter factors. For each candidate filter factor generated, the smoothed two-dimensional path corresponding to the filter factor is projected onto a three-dimensional surface to obtain the platform processing path. ,in Provide the 3D coordinates of the processing path points for the platform. Let N be the normal vector of the platform processing path point, and N be the number of path points in the platform processing path.

[0028] Calculate the defocusing amount, laser incident angle, and galvanometer scanning range, and verify the constraints. Take the candidate filter factor that does not meet the constraints for the first time as the upper bound of the filter factor. Determine the search range .

[0029] Furthermore, in step S6, for each midpoint filter factor Perform iterations and determine constraint feasibility: if feasible, then let If that is not feasible, then order ;

[0030] In step S7, the search is terminated and the maximum feasible filter factor is output when any of the following conditions are met:

[0031] The convergence criterion for the filter factor is Or reach the maximum number of iterations ,in, The convergence threshold of the filter factor. To prevent small amounts from being divided by zero.

[0032] Furthermore, in step S3, the segment filter factor weight matrix is ​​initialized. Let the smooth two-dimensional path of the 0th iteration be ;

[0033] For any number of iterations Construct a diagonal filter factor weight matrix based on the current segment filter factor weight, and solve for the weighted filtering result.

[0034] Furthermore, in step S5:

[0035] Construction point constraint occupancy rate With point margin :

[0036] , ;

[0037] Aggregate point margins into smoothing operator segments:

[0038] ;

[0039] From continuous monotonic functions Mapped to segment filter factor weights:

[0040] ,

[0041] ,

[0042] in, Used to control the steepness of the transition. Used to control the upper and lower bounds of the filter factor weights in the control segment;

[0043] Weighted smoothing is used to suppress oscillations:

[0044] ,

[0045] in, It is obtained by mapping the constraint margins through k iterations. This is the weighted smoothing coefficient. It is the segment filter factor weight between two adjacent iterations.

[0046] Furthermore, in step S7:

[0047] When the path convergence criterion is met, the iteration is considered converged and the iteration is terminated.

[0048] ,

[0049] in, This is the path convergence threshold. , A smooth two-dimensional path between two adjacent iterations;

[0050] If the number of weight iterations reaches the preset limit If the convergence criterion is still not met, the iteration terminates and the current iteration is output. and The solution result is used as the corresponding candidate filter factor.

[0051] Furthermore, in step 8:

[0052] The multi-axis laser flying processing system has a base, with the laser emitter, galvanometer, and motion platform mounted on the base respectively. The workpiece is located at the execution end of the motion platform, and the U and V axes of the galvanometer are in the same direction as the X and Y axes of the motion platform.

[0053] The difference between the X and Y axis components of the final platform machining path and the target machining path can be used to obtain the compensation motion commands for the U and V axes of the galvanometer, thus generating the galvanometer machining path.

[0054] Galvanometer processing path R:

[0055] ,

[0056] ,

[0057] in, , The X and Y axis components of the target machining path point i. , The platform processing path point i has X and Y axis components;

[0058] Galvanometer processing range:

[0059] ;

[0060] .

[0061] Furthermore, in step 8:

[0062] The multi-axis laser flying processing system has a base, with the laser emitter, workpiece and motion platform respectively positioned on the base, and the galvanometer located at the execution end of the motion platform. The galvanometer focal plane coordinate system {G} changes in real time with the motion of the execution end.

[0063] During the processing, the laser beam is always with In the same direction, define the residual vector in the workpiece coordinate system {W}:

[0064] ;

[0065] The residual is decomposed into the following within the focal plane of the galvanometer:

[0066] ;

[0067] Based on the attitude transformation matrix of the workpiece coordinate system {W} and the galvanometer focal plane coordinate system {G} get ,Pick and The components on are used as galvanometer commands:

[0068] ,

[0069] The set of galvanometer instructions along the final platform machining path constitutes the galvanometer machining path R.

[0070] The technical solution provided by this invention may include the following beneficial effects:

[0071] (1) A set of hard constraints is constructed using the defocus amount, laser incident angle, and galvanometer scanning range. Projection calculations and constraint evaluations are performed on the path after each filtering, and the segment filtering factor weights are updated and the filtering factor is searched accordingly to achieve closed-loop optimization. Compared with the single global filtering method, this invention achieves higher smoothness of the platform processing path in the constraint-relaxed region through adaptive adjustment of the segment filtering factor weights based on constraint feedback and search for the maximum feasible filtering factor, thereby reducing the platform processing path length and impact vibration, and improving the efficiency of linkage processing.

[0072] (2) The constraint occupancy rate and margin are aggregated into the smoothing operator segment and mapped to the segment filter factor weight through a continuous monotonic function to achieve adaptive allocation of the path smoothness; at the same time, weight smoothing is introduced to suppress iterative oscillations and improve the convergence stability of the segment filter factor weight. This invention calculates the constraint margin point by point and adaptively adjusts the segment filter factor weight, so that the tense region does not go out of bounds and the smoothness continues to increase in the relaxed region, breaking through the conservative bottleneck caused by the traditional global unified penalty.

[0073] (3) Perform inner-layer iterations under a given filter factor until convergence or the upper limit is reached; determine the infeasible upper bound by doubling in the outer layer and search for the maximum feasible filter factor in the logarithmic domain, so as to obtain the greatest possible smoothness under the premise of satisfying the constraints.

[0074] (4) This invention does not rely on a fixed filter or a fixed mathematical form: the specific functions of smoothing operator, deviation cost and smoothing penalty can be replaced. As long as the core idea of ​​three-constraint closed-loop evaluation and constraint margin driven adaptive filtering is maintained, the same technical effect can be achieved, making the scope of application of this invention wide. Attached Figure Description

[0075] Figure 1 is a schematic diagram of the defocusing amount, incident angle, and galvanometer scanning range of laser multi-axis flying processing;

[0076] Figure 2 is a flowchart of the laser multi-axis flight processing path planning method with adaptive weighted filtering according to the present invention;

[0077] Figure 3 is a schematic diagram of the coordinate system of the first multi-axis laser flying processing system;

[0078] Figure 4 is a schematic diagram of the coordinate system of the second type of multi-axis laser flying processing system;

[0079] Figure 5 is a three-dimensional comparison of the target processing path, the platform processing path obtained by the single global filtering of the background technology, and the platform processing path obtained by the adaptive weighted filtering of the present invention.

[0080] Figure 6 is a planar comparison diagram of the target processing path, the platform processing path obtained by the single global filtering of the background technology, and the platform processing path obtained by the adaptive weighted filtering of the present invention. Detailed Implementation

[0081] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that the specific embodiments described herein are for illustrative and explanatory purposes only and are not intended to limit the present invention.

[0082] Referring to Figure 2, an adaptive weighted filtering laser multi-axis flight processing path planning method according to an embodiment of the present invention includes the following steps:

[0083] S1: Input the 3D surface model of the workpiece and the target 2D path in the workpiece coordinate system, and set the constraint thresholds for defocusing amount, laser incident angle, and galvanometer scanning range as follows: , , Set the lower bound of the filter factor λ. ;

[0084] S2: Determine the search interval for the filter factor that satisfies the constraints of defocusing amount, laser incident angle, and galvanometer scanning range. ;

[0085] S3: Based on the filter factor search interval The midpoint filter factor is generated using the logarithmic domain bisection method. Initialize the weight matrix of the filter factors;

[0086] S4: Based on midpoint filter factor The weight matrix of the initial filtering factor is used to perform weighted filtering on the target two-dimensional path F through the adaptive weighted filtering objective function to obtain the smooth two-dimensional path E;

[0087] Adaptive weighted filtering objective function:

[0088] ,

[0089] in, For the points on the target two-dimensional path, For points on a smooth 2D path, This indicates a deviation from the cost. Indicates a smooth penalty. λ represents the segment filter factor weight in the weight matrix W, L represents the smoothing operator, and λ represents the filter factor. The larger the value, the smoother the result, but the easier it is to deviate from the original path.

[0090] S5: Project the smooth 2D path onto the 3D surface to obtain the platform machining path, and calculate the constraint parameters and constraint margins. The constraint parameter is the defocus amount. laser incident angle and galvanometer scanning range Generate segment filter factor weights based on constraint margins Update the weight matrix; repeat steps S3 and S4 until the path converges or the maximum number of weight iterations is reached, and output the final platform processing path;

[0091] S6: Determine whether the platform processing path output by S5 meets the constraints. If it does, then... Updated to Otherwise Updated to ;

[0092] S7: Repeat S3-S6 until the filter factor converges or the maximum number of iterations of the filter factor is reached, and output the final platform processing path;

[0093] S5: Based on the target machining path and the final platform machining path, the galvanometer machining path is calculated. (Machining path on the focal plane).

[0094] In this scheme, under the premise that the three types of constraints of defocus amount, incident angle and galvanometer scanning range are satisfied at the same time, the weight of the segment filter factor is adaptively adjusted through constraint feedback to achieve adaptive allocation of the path smoothness. In the constrained tense area, the smoothness is automatically reduced to ensure that it does not go out of bounds; in the constrained loose area, the smoothness is automatically increased to shorten the platform processing path length as much as possible, reduce curvature and vibration impact, thereby improving the average feed speed and processing efficiency of the motion platform.

[0095] In this scheme, the target two-dimensional path corresponds to the planar pattern, the smooth two-dimensional path is the two-dimensional path output by filtering, the target machining path is the projection result of the target two-dimensional path on the three-dimensional curved surface, the platform machining path is the projection result of the smooth two-dimensional path on the three-dimensional curved surface, and the galvanometer machining path is the machining path on the focal plane.

[0096] Preferably, in step S4, For squared, robust loss functions, or absolute values; It is a squared or robust loss function, where L is the first-order difference, second-order difference, or spline second derivative. The specific functions of deviation cost, smoothing penalty, and smoothing operator can be replaced, making this scheme independent of a fixed filter or fixed mathematical form. To ensure the convexity of the objective function and the stability of the outer parameter search, and Preferred convex functions, A linear difference operator is preferred. The adaptive weighted filtering solution process of this invention includes inner-layer segment filter factor weight iteration and outer-layer filter factor λ iterative search.

[0097] Referring to Figure 1, specifically, in step S1:

[0098] Defocus amount is defined as the target processing path point Platform processing path points Unit normal vector at the platform processing path point Absolute distance in direction: defocus amount Platform processing path point unit normal vector Correspondingly, in the workpiece machining coordinate system, , .

[0099] In curved surface laser flight processing, the galvanometer lens must be a telecentric lens to ensure that the laser beam is aligned with the normal vector of the platform's processing path. Parallel, the laser incident angle is defined as the normal vector of the target machining path point. platform processing path point normal vector The included angle, laser incident angle ;

[0100] The galvanometer scanning range is the maximum millimeter displacement of the U-axis and V-axis in the galvanometer focal plane coordinate system. , It's understandable. , It is the millimeter displacement in the coordinate system of the galvanometer focal plane;

[0101] The constraints on defocusing amount, laser incident angle, and galvanometer scanning range are summarized as follows:

[0102] .

[0103] As an example, in the workpiece coordinate system {W}, the target two-dimensional path Constraint threshold: Maximum allowable defocus amount (mm), maximum permissible laser incident angle (degrees / °), maximum permissible galvanometer scanning range (mm).

[0104] This invention aims to obtain the largest possible filter factor λ while satisfying constraints, thereby achieving a higher degree of smoothing and a shorter processing time. Specifically, in step S2:

[0105] From the lower bound of the filter factor The process begins by multiplying the generation of candidate filter factors. For each candidate filter factor generated, the smoothed two-dimensional path corresponding to that filter factor is projected onto a three-dimensional surface to obtain the platform machining path. ,in Provide the 3D coordinates of the processing path points for the platform. Let N be the normal vector of the platform processing path point, and N be the number of path points in the platform processing path.

[0106] Calculate the defocusing amount, laser incident angle, and galvanometer scanning range, and verify the constraints. Take the candidate filter factor that does not meet the constraints for the first time as the upper bound of the filter factor. Determine the search range .

[0107] Furthermore, in step S6, for each midpoint filter factor Perform iterations and determine constraint feasibility: if feasible, then let If that is not feasible, then order ;

[0108] In the search range Internal search for the maximum feasible filter factor :

[0109] ;

[0110] In step S7, the search is terminated and the maximum feasible filter factor is output when any of the following conditions are met:

[0111] The convergence criterion for the filter factor is (The relative width of the interval converges) or the maximum number of iterations is reached. (Maximum iteration count protection), where, The convergence threshold of the filter factor. To prevent small amounts from being divided by zero.

[0112] In step S3, the segment filter factor weight matrix is ​​initialized. Let the smooth two-dimensional path of the 0th iteration be ;

[0113] For any number of iterations Construct a diagonal filter factor weight matrix based on the current segment filter factor weight, and solve for the weighted filtering result.

[0114] Furthermore, in step S5:

[0115] Construction point constraint occupancy rate With point margin :

[0116] , ;

[0117] Aggregate point margins into smoothing operator segments:

[0118] ;

[0119] From continuous monotonic functions Mapped to segment filter factor weights:

[0120] ,

[0121] ,

[0122] in, Used to control the steepness of the transition. Used to control the upper and lower bounds of the filter factor weights in the control segment;

[0123] Weighted smoothing is used to suppress oscillations:

[0124] ,

[0125] in, It is obtained by mapping the constraint margins through k iterations. This is the weighted smoothing coefficient. It is the segment filter factor weight between two adjacent iterations.

[0126] When a segment approaches the constraint boundary (small margin), the weight of the segment filter factor tends to be smaller, which weakens the penalty of the smoothing operator for that segment and avoids excessive smoothing that causes the defocus amount, laser incident angle, and galvanometer scanning range to exceed the constraints.

[0127] When a segment is far from the constraint boundary (with a large margin), the segment filtering factor weight tends to be larger, making the segment smoother and increasing the upper limit of the feed rate that the segment can reach.

[0128] Furthermore, in step S7:

[0129] When the path convergence criterion is met, the iteration is considered converged and the iteration is terminated.

[0130] ,

[0131] in, This is the path convergence threshold. , A smooth two-dimensional path between two adjacent iterations;

[0132] If the number of weight iterations reaches the preset limit If the convergence criterion is still not met, the iteration terminates and the current iteration is output. and The solution results are used as the corresponding candidate filter factors to ensure that the algorithm can terminate in a finite number of steps.

[0133] For multi-axis laser processing systems with different structures, the methods for solving the galvanometer processing path are not entirely the same. The core logic is to solve the galvanometer processing path based on the target processing path and the platform processing path.

[0134] Specifically, referring to Figure 3, when using the first multi-axis laser flying processing system, in step 5:

[0135] The multi-axis laser flying processing system has a base, and the laser emitter, galvanometer and motion platform are respectively mounted on the base. The workpiece is located at the execution end of the motion platform. The motion platform is a five-axis motion platform, and the U and V axes of the galvanometer are in the same direction as the X and Y axes of the motion platform. Then the U and V axes of the galvanometer can be used as micro-compensation platforms for the X and Y axes of the motion platform to compensate for the motion deviation of the X and Y axes of the motion platform during the processing.

[0136] The difference between the X and Y axis components of the final platform machining path and the target machining path can be used to obtain the compensation motion commands for the U and V axes of the galvanometer, thus generating the galvanometer machining path.

[0137] Galvanometer processing path R:

[0138] ,

[0139] ,

[0140] in, , The X and Y axis components of the target machining path point i. , The platform processing path point i has X and Y axis components;

[0141] Galvanometer processing range:

[0142] ;

[0143] .

[0144] In the second type of multi-axis laser flying processing system, in step 8:

[0145] The multi-axis laser flying machining system has a base, with the laser emitter, workpiece, and motion platform positioned on the base. A galvanometer is located at the end effector of the motion platform, which is a five-axis platform. The galvanometer deflects synchronously with the attitude of the end effector, and the focal plane coordinate system {G} of the galvanometer changes in real time with the motion of the end effector. Since the motion platform always moves along the machining path of the workpiece coordinate system platform, the laser beam is always aligned with the workpiece during the machining process. In the same direction, define the residual vector in the workpiece coordinate system {W}: ;

[0146] The residual is decomposed into the following within the focal plane of the galvanometer: ;

[0147] Based on the attitude transformation matrix of the workpiece coordinate system {W} and the galvanometer focal plane coordinate system {G} get ,Pick (Positive U-axis) and The component on the positive V-axis is used as the galvanometer command:

[0148] ,

[0149] The set of galvanometer instructions along the final platform machining path constitutes the galvanometer machining path R.

[0150] In a multi-axis laser flight processing system, the end effector of the motion platform refers to its execution / output end. It is the mechanical structure that directly mounts the laser processing head and galvanometer assembly, and it is also the final carrier for the motion platform's motion accuracy and attitude. The motion platform achieves position and attitude adjustment through multi-axis linkage. The galvanometer is not installed independently but is directly fixed to the end effector structure of the motion platform. When the motion platform completes translational, deflective, or other movements, the attitude of its end effector changes synchronously, and the spatial attitude of the galvanometer also changes accordingly.

[0151] In this embodiment of the invention, take and Let L be the square of the difference matrix. Construct an adaptive weighted filtering function and take the segmented filter factor weights. , Weighted smoothing coefficient Steepness coefficient Multiplication factor , , The experimental equipment structure adopts the first type of multi-axis laser flying processing system described above, and the final platform processing path is shown in Figure 3. As can be seen from Figures 5 and 6, the platform processing path obtained by this invention is smoother in the corner areas than the platform processing path obtained by the background technology method. Finally, through experimental verification, compared with the platform processing path obtained by the single global filtering of the background technology, the processing efficiency of the platform processing path provided by this invention is improved by 9.2%.

[0152] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0153] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0154] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, optical disk) and includes several instructions to cause a multimedia terminal device (which may be a mobile phone, computer, television receiver, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0155] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A laser multi-axis flight processing path planning method with adaptive weighted filtering, characterized in that, The steps include: S1: Input the 3D surface model of the workpiece and the target 2D path in the workpiece coordinate system, and set the constraint thresholds for defocusing amount, laser incident angle, and galvanometer processing range as follows: 、 、 Set the lower bound of the filter factor λ. S2: Determine the search interval for the filter factor that satisfies the constraints of defocusing amount, laser incident angle, and galvanometer processing range. S3: Search interval based on filter factor The midpoint filter factor is generated using the logarithmic domain bisection method. S4: Initialize the weight matrix of the filter factor; The weight matrix of the initialized filter factors is used to perform weighted filtering on the target two-dimensional path F to obtain a smooth two-dimensional path E through an adaptive weighted filtering objective function; the adaptive weighted filtering objective function is: ,in, For the points on the target two-dimensional path, For points on a smooth 2D path, This indicates a deviation from the cost. Indicates a smooth penalty. S5: The segment filter factor weights in the weight matrix W are represented by L, and the smoothing operator is represented by L. S6: The smoothed two-dimensional path is projected onto the three-dimensional surface to obtain the platform machining path, and the constraint parameters and constraint margins are calculated. The constraint parameter is the defocus amount. laser incident angle And galvanometer processing range Generate segment filter factor weights based on constraint margins Update the weight matrix; repeat steps S3 and S4 until the path converges or the maximum number of weight iterations is reached, and output the platform processing path; S6: Determine whether the platform processing path output in S5 satisfies the constraints. If it does, then... Updated to Otherwise Updated to S7: Repeat S3-S6 until the filter factor converges or reaches the maximum number of iterations, and output the final platform processing path; S8: Calculate the galvanometer processing path based on the target processing path and the final platform processing path. In step S5: constructing point constraint occupancy rate With point margin ： ， Aggregate point margins into smoothing operator segments: ; by continuous monotonic functions Mapped to segment filter factor weights: ， ,in, Used to control the steepness of the transition. Used to control the upper and lower bounds of the filter factor weights in the control segment; weight smoothing is employed to suppress oscillations. ,in, It is obtained by mapping the constraint margins through k iterations. This is the weighted smoothing coefficient. It is the segment filter factor weight between two adjacent iterations.

2. The method according to claim 1, characterized in that, In step S4 For squared, robust loss functions, or absolute values; It is the squared or robust loss function, and L is the first-order difference, second-order difference, or spline second derivative.

3. The method according to claim 1, characterized in that, In step S1: the defocusing amount is defined as the target processing path point. Platform processing path points Unit normal vector at the platform processing path point Absolute distance in direction: defocus amount The laser incident angle is defined as the normal vector of the target machining path point. platform processing path point normal vector The included angle, laser incident angle The galvanometer machining range is the maximum millimeter displacement of the U-axis and V-axis in the galvanometer focal plane coordinate system. ， The constraints of defocusing amount, laser incident angle, and galvanometer processing range must be satisfied simultaneously: 。 4. The method according to claim 3, characterized in that, In step S2: from the lower bound of the filter factor The process begins by multiplying candidate filter factors. For each candidate filter factor generated, the smoothed two-dimensional path corresponding to the filter factor is projected onto a three-dimensional surface to obtain the platform processing path. ,in Provide the 3D coordinates of the processing path points for the platform. Let N be the normal vector of the platform machining path points, and N be the number of path points on the platform machining path. Calculate the defocusing amount, laser incident angle, and galvanometer machining range, and verify the constraints. Take the candidate filter factor that first fails to meet the constraints as the upper bound of the filter factor. Determine the search range 。 5. The method according to claim 1, characterized in that, In step S6, the midpoint filter factor Perform iterations and determine constraint feasibility: if feasible, then let If that is not feasible, then order In step S7, the search is terminated and the maximum feasible filter factor is output when any of the following conditions are met: the filter factor convergence criterion is... Or reach the maximum number of iterations ,in, The convergence threshold of the filter factor. To prevent small amounts from being divided by zero.

6. The method according to claim 3, characterized in that, In step S3, the segment filter factor weight matrix is ​​initialized. Let the smooth two-dimensional path of the 0th iteration be ； For any number of iterations Construct a diagonal filter factor weight matrix based on the current segment filter factor weight, and solve for the weighted filtering result.

7. The method according to claim 1, characterized in that, In step S7: when the path convergence criterion is met, the iteration is considered to have converged and the iteration is terminated. ,in, This is the path convergence threshold. 、 For a smooth two-dimensional path between two consecutive iterations; if the number of weight iterations reaches a preset upper limit. If the convergence criterion is still not met, the iteration terminates and the current iteration is output. and The solution result is used as the corresponding candidate filter factor.

8. The method according to claim 3, characterized in that, In step S8: the multi-axis laser flying processing system has a base, and the laser emitter, galvanometer, and motion platform are respectively mounted on the base. The workpiece is located at the execution end of the motion platform, and the U and V axes of the galvanometer are in the same direction as the X and Y axes of the motion platform. The difference between the X and Y axis components of the final platform processing path and the target processing path yields the compensation motion commands for the U and V axes of the galvanometer, generating the galvanometer processing path; the galvanometer processing path R: ， ,in, 、 The X and Y axis components of the target machining path point i. 、 Platform machining path point i in the X and Y axis components; galvanometer machining range: ； 。 9. The method according to claim 3, characterized in that, In step S8: the multi-axis laser flying processing system has a base, the laser emitter, the workpiece, and the motion platform are respectively positioned on the base, the galvanometer is located at the execution end of the motion platform, and the focal plane coordinate system {G} of the galvanometer changes in real time with the movement of the execution end; during the processing, the laser beam is always aligned with the laser platform. In the same direction, define the residual vector in the workpiece coordinate system {W}: The residual is decomposed into the following components within the focal plane of the galvanometer: Based on the attitude transformation matrix of the workpiece coordinate system {W} and the galvanometer focal plane coordinate system {G} get ,Pick and The components on are used as galvanometer commands: The set of galvanometer instructions along the platform machining path constitutes the galvanometer machining path R.

Images