A method and system for automatic laser cleaning trajectory planning of irregularly shaped parts

Abstract

This invention provides a method and system for automatic laser cleaning trajectory planning of irregularly shaped parts, relating to the field of robot control. The method includes: visually scanning the irregularly shaped parts; planning a cleaning trajectory based on the point cloud coordinates on the surface of the part's 3D model, with the goal of reducing the robot arm's running time, energy consumption, impact, and acceleration; detecting and eliminating interference at the trajectory points on the determined cleaning trajectory; converting the coordinates of the trajectory points on the cleaning trajectory into rotation angles of each joint of the robot arm using a robot arm kinematic model; optimizing the rotation angles of each joint of the robot arm using a Jacobian matrix combined with iterative solution, with the goal of reducing the difference between the robot arm's end effector and the trajectory points on the cleaning trajectory; controlling the robot arm to move according to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angles of each joint of the robot arm, and using laser to automatically clean the irregularly shaped parts.

Description

Technical Field

[0001] This invention relates to the field of trajectory control technology, and in particular to a method and system for automatic laser cleaning trajectory planning for irregularly shaped parts. Background Technology

[0002] Laser cleaning uses a focused laser beam to precisely clean specific areas on the surface of parts, avoiding damage to areas that do not require cleaning. Furthermore, laser cleaning is fast, making it suitable for rapid cleaning of parts in large-scale production, significantly reducing cleaning time.

[0003] Currently, automatic laser cleaning mainly relies on operators to control robotic arms to perform indiscriminate scanning from left to right and from top to bottom. This only allows for simple planar scanning and cleaning trajectory planning, resulting in low utilization of the robotic arms. It only performs vertical downward scanning and completely abandons the angular rotation function of the robotic arms.

[0004] Furthermore, for workpieces with complex surfaces, the cumbersome process of manual programming by operators is greatly increased, and the programming operation has a large error, affecting the accuracy of the laser automatic cleaning trajectory and the precision of the laser automatic cleaning. Summary of the Invention

[0005] To address the current technical issues of relying primarily on operators to control robotic arms for indiscriminate scanning from left to right and top to bottom, which only allows for simple planar scanning cleaning trajectory planning and results in low utilization of the robotic arms, and significantly increases the cumbersomeness of manual programming for complex workpiece surfaces, leading to substantial errors in the programming operation and affecting the accuracy and precision of automatic laser cleaning trajectories, this invention provides a method and system for automatic laser cleaning trajectory planning for irregularly shaped parts.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] First aspect:

[0008] This invention provides a method for automatic laser cleaning trajectory planning for irregularly shaped parts, comprising:

[0009] S1: Perform visual scanning on the irregularly shaped parts to generate a three-dimensional model of the parts;

[0010] S2: Based on the point cloud coordinates on the surface of the three-dimensional model of the component, with the goal of reducing the running time, energy consumption, impact and acceleration of the robotic arm, a cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm.

[0011] S3: Detect and eliminate interference at the trajectory points on the determined cleaning trajectory;

[0012] S4: Construct the kinematic model of the robotic arm;

[0013] S5: Using the kinematic model of the robotic arm, the coordinates of the trajectory points on the cleaning trajectory are converted into the rotation angles of each joint of the robotic arm;

[0014] S6: With the goal of reducing the difference between the end effector of the robotic arm and the trajectory points on the cleaning trajectory, the rotation angles of each joint of the robotic arm are optimized by using the Jacobian matrix in combination with iterative solution.

[0015] S7: According to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angle of each joint of the robotic arm, control the movement of the robotic arm and use laser to automatically clean irregularly shaped parts.

[0016] The second aspect:

[0017] This invention provides a laser automatic cleaning trajectory planning system for irregularly shaped parts, comprising:

[0018] processor;

[0019] The memory stores computer-readable instructions, which, when executed by the processor, implement the laser automatic cleaning trajectory planning method for irregularly shaped parts as described in the first aspect.

[0020] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0021] In this invention, the irregularly shaped parts are visually scanned to generate a three-dimensional model of the parts. With the goal of reducing the running time, energy consumption, impact, and acceleration of the robotic arm, an automatic cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm to improve the accuracy of the automatic cleaning trajectory. Then, according to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angles of each joint of the robotic arm, the movement of the robotic arm is controlled, and laser is used to automatically clean the irregularly shaped parts, thereby improving the utilization rate of the robotic arm and improving the efficiency and accuracy of laser automatic cleaning. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating a laser automatic cleaning trajectory planning method for irregularly shaped parts provided in an embodiment of the present invention;

[0024] Figure 2 This is a schematic diagram of a laser automatic cleaning trajectory planning system for irregularly shaped parts, provided in an embodiment of the present invention. Detailed Implementation

[0025] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0026] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0027] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0028] Reference manual attached Figure 1 The diagram shows a flowchart of a laser automatic cleaning trajectory planning method for irregularly shaped parts provided by an embodiment of the present invention.

[0029] This invention provides a method for automatic laser cleaning trajectory planning for irregularly shaped parts, comprising the following steps:

[0030] S1: Perform visual scanning on irregularly shaped parts to generate 3D models of the parts.

[0031] In one possible implementation, S1 specifically involves: visually scanning the irregularly shaped parts using a high-precision camera or reverse modeling equipment to generate a three-dimensional model of the parts.

[0032] In this invention, generating three-dimensional models of parts using a high-precision camera or reverse modeling equipment not only provides high-quality model data to support precise planning of subsequent laser cleaning trajectories, but also significantly improves work efficiency and reduces trial-and-error costs, which is an important prerequisite for achieving automated and precise cleaning.

[0033] S2: Based on the point cloud coordinates on the surface of the 3D model of the parts, with the goal of reducing the running time, energy consumption, impact and acceleration of the robotic arm, the cleaning trajectory is planned by combining flower pollination with genetic optimization algorithm.

[0034] The flower pollination algorithm combined with the genetic optimization algorithm combines the Flower Pollination Algorithm (FPA) and the Genetic Algorithm (GA). The flower pollination algorithm is an optimization algorithm based on the pollination behavior of flowers in nature, simulating the process by which plants propagate their population through the spread of pollen via wind, insects, and other mediums. The genetic algorithm is a stochastic search algorithm based on the principles of natural selection and genetics, simulating the process of biological evolution. Through genetic operations (selection, crossover, mutation), it continuously optimizes the population to find a near-optimal solution to the problem. While the genetic algorithm excels at global search, it lacks local search capabilities. Introducing the flower pollination algorithm allows the algorithm to flexibly adjust its search strategy according to the complexity of the problem and the optimization stage.

[0035] In one possible implementation, S2 specifically includes sub-steps S201 to S203:

[0036] S201: Construct an optimization objective function with the goal of reducing the robotic arm's running time, energy consumption, impact, and acceleration.

[0037] Optionally, the objective function to be optimized is as follows:

[0038]

[0039] Wherein, min represents minimization, F represents the objective function, c1 represents the running time of the robotic arm, c2 represents the energy consumption of the robotic arm, c3 represents the impact on the robotic arm joints, c4 represents the acceleration of the robotic arm joints, λ1 represents the weighting coefficient of the running time of the robotic arm, λ2 represents the weighting coefficient of the energy consumption of the robotic arm, λ3 represents the weighting coefficient of the impact on the robotic arm joints, λ4 represents the weighting coefficient of the acceleration of the robotic arm joints, ξ1 represents the normalization coefficient of the running time of the robotic arm, ξ2 represents the normalization coefficient of the energy consumption of the robotic arm, ξ3 represents the normalization coefficient of the impact on the robotic arm joints, and ξ4 represents the normalization coefficient of the acceleration of the robotic arm joints.

[0040] Those skilled in the art can set the weighting coefficients λ1 for the running time of the robotic arm, λ2 for the energy consumption of the robotic arm, λ3 for the impact of the robotic arm joints, and λ4 for the acceleration of the robotic arm joints according to the actual situation. This invention does not limit these settings.

[0041] In this invention, an optimization objective function based on running time, energy consumption, impact, and acceleration is constructed, making the laser cleaning trajectory planning of the robotic arm more scientific, flexible, and efficient. This method not only significantly improves cleaning quality and system reliability but also reduces operating costs and environmental impact, meeting the core requirements of modern intelligent manufacturing.

[0042] Optionally, the operating time of the robotic arm is as follows:

[0043]

[0044] Where Tk represents the running time on the k-th cleaning trajectory.

[0045] Optionally, the energy consumption of the robotic arm is as follows:

[0046]

[0047] Among them, E ki I represents the energy consumption of the i-th joint on the k-th segment of the cleaning trajectory. i Let ω represent the moment of inertia of the i-th joint. ki P represents the angular velocity of the i-th joint on the k-th segment of the cleaning trajectory. k+1,i P represents the rotation angle of the i-th joint on the (k+1)-th segment of the cleaning trajectory. ki T represents the rotation angle of the i-th joint on the k-th cleaning trajectory. ki Let represent the running time of the i-th joint on the k-th segment of the cleaning trajectory, and n represent the total number of trajectory points.

[0048] Optionally, the impact on the robotic arm joints is specifically as follows:

[0049]

[0050] Where, γ ki α represents the angular impact of the i-th joint on the k-th segment of the cleaning trajectory. ki Let α represent the angular acceleration of the i-th joint on the k-th segment of the cleaning trajectory. k-1,i It represents the angular acceleration of the i-th joint on the (k-1)-th segment of the cleaning trajectory.

[0051] Optionally, the acceleration of the robotic arm joints is specifically as follows:

[0052]

[0053] S202: Set constraints in the cleaning trajectory planning process.

[0054] Optionally, the constraints are as follows:

[0055]

[0056] Where λ1 represents the weighting coefficient of the robotic arm's running time, λ2 represents the weighting coefficient of the robotic arm's energy consumption, λ3 represents the weighting coefficient of the impact on the robotic arm's joints, λ4 represents the weighting coefficient of the acceleration of the robotic arm's joints, and ω kiLet τi represent the angular velocity of the i-th joint on the k-th segment of the cleaning trajectory, and τ1 represent the velocity safety factor. i J represents the maximum velocity of the i-th joint. i Let α represent the moment of inertia of the i-th joint. ki Let F represent the angular acceleration of the i-th joint on the k-th segment of the cleaning trajectory, τ2 represent the driving torque safety factor, and F represent the angular acceleration of the i-th joint on the k-th segment of the cleaning trajectory. i This represents the maximum driving torque of the i-th joint.

[0057] It should be noted that limiting the angular velocity of the joints prevents the robotic arm from moving too fast, which could lead to trajectory instability or hardware damage. Furthermore, limiting the joint drive torque prevents hardware overload due to excessive acceleration.

[0058] In this invention, by setting reasonable constraints, efficiency, safety, and hardware protection can be effectively balanced during trajectory optimization, ensuring that the trajectory planning results not only have optimal performance but also execute stably and safely in actual operation. This method meets the needs of industrial applications and provides important support for the efficient and reliable laser cleaning of robotic arms.

[0059] S203: Under the constraints of the conditions, with the goal of minimizing the optimization objective function, the cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm.

[0060] Specifically, the objective function can be used as the fitness function of the flower pollination combined with the genetic optimization algorithm.

[0061] Initialize the population, which contains multiple individuals. Each individual represents a feasible cleaning trajectory scheme. Each individual has multiple components, and each component represents the coordinates of a trajectory point.

[0062] An elite selection strategy was adopted to remove the 30% of individuals with the lowest fitness values ​​to form the first population.

[0063] In this invention, an elite selection strategy is adopted, removing the bottom 30% of individuals by fitness value. This significantly improves population quality. By retaining individuals with higher fitness, the population is ensured to converge towards optimal solutions, accelerating the optimization process. Simultaneously, eliminating individuals with lower fitness reduces interference from invalid solutions and lowers computational costs. Furthermore, this strategy balances population diversity and convergence speed, providing a better genetic basis for subsequent crossover and mutation operations, thereby improving optimization efficiency and global solution search capabilities.

[0064] Crossover operations are performed on individuals in the first group to form the second group:

[0065] Y1 = rand × X1 + (1 - rand) × X2

[0066] Y2=rand×X2+(1-rand)×X1

[0067] Where Y1 and Y2 represent new individuals, X1 represents the first parent, X2 represents the second parent, and rand represents a random number between 0 and 1.

[0068] In this invention, a crossover operation is performed on individuals in the first population. By combining the genes of two parent individuals with random weights to generate two new individuals, it helps to retain superior genes while exploring new solution spaces. This enhances population diversity, prevents the algorithm from getting stuck in local optima, and accelerates the search process for the global optimum, providing higher-quality candidate solutions for subsequent optimization stages.

[0069] Mutation operations are performed on individuals in the second population to form the third population:

[0070]

[0071] Where Y3 represents the new individual, X3 represents the parent entity, and X... max X represents the individual with the highest fitness value. min This represents the individual with the smallest fitness value, and rand represents a random number between 0 and 1.

[0072] In this invention, mutation operations are performed on individuals in the second population. This introduces diversity while preserving high-quality solutions, preventing the population from prematurely converging to local optima. Furthermore, this mutation method guides the direction of mutation by comparing the largest and smallest individuals, which helps to explore the solution space more efficiently, improves global search capabilities, and provides a better gene pool for subsequent optimization.

[0073] Individuals in the third population are updated in position using a flower pollination optimization algorithm to form the fourth population. A random number is generated, and the transition probability is checked against the random number. If the probability is greater than the random number, cross-pollination is performed. Otherwise, self-pollination is performed.

[0074] When cross-pollination occurs, the individual's position is updated according to the Levi flight mechanism:

[0075]

[0076] in, This represents the position of the i-th individual in the (t+1)-th iteration. Let θ represent the position of the i-th individual in the t-th iteration. t Let L represent the step size influence factor at the t-th iteration, where L represents the step size. This represents the global optimal solution at the t-th iteration.

[0077]

[0078] Where Γ represents the standard Gamma function, λ represents the exponential parameter, and s represents the scale parameter.

[0079] In this invention, during cross-pollination, the Lévy flight mechanism is used to update individual positions. By introducing a step size influence factor and a step size itself, individuals are guided closer to the global optimum. Simultaneously, the step size generated by the Lévy distribution provides randomness and long-distance jump capability. This enhances the algorithm's global search capability, particularly by effectively exploring potential optimal regions in the solution space and avoiding getting trapped in local optima. Furthermore, the randomized nature of the Lévy flight's long and short step sizes balances the search range with refined search, improving the algorithm's optimization efficiency and robustness.

[0080] Optionally, the step size influence factor is as follows:

[0081]

[0082] Where, θ t represents the step size influence factor at the t-th iteration, q represents the scaling factor, and T represents the maximum number of iterations.

[0083] In this invention, the step size influence factor allows the step size to gradually decrease as the number of iterations increases. A larger step size in the initial stage helps the algorithm quickly explore the solution space and improves global search capability; a smaller step size in the later stages allows the algorithm to perform fine-tuning in local regions of the solution space, improving the accuracy and convergence performance of the solution.

[0084] When self-pollination occurs, the individual's position is updated according to the golden sine mechanism:

[0085]

[0086] x1 = -π + 2π(1 - τ)

[0087] x2=-π+2πτ

[0088] in, This represents the position of the randomly selected j-th individual in the t-th iteration. Let r1 and r2 represent the position of the kth randomly selected individual in the tth iteration, r1 and r2 represent random numbers between 0 and 2π, x1 and x2 represent the self-pollination coefficients, and τ represent the golden ratio.

[0089] In this invention, a golden sine mechanism is used to update individual positions during self-pollination, making the update both exploratory and stable. The advantage of this approach is that it leverages the balanced properties of the sine function and the golden ratio to generate smoother and finer update paths, enhancing local search capabilities while preventing the algorithm from getting trapped in local optima. Furthermore, randomly selecting individuals based on their differences further promotes population diversity, contributing to improved optimization efficiency and solution quality.

[0090] The third and fourth populations were merged to form the fifth population.

[0091] Determine if the current iteration count has reached the maximum iteration count; if so, output the cleaning trajectory scheme represented by the individual with the highest fitness in the fifth population; otherwise, return to continue iterating.

[0092] In this invention, a cleaning trajectory planning method combining flower pollination and a genetic optimization algorithm leverages the advantages of both algorithms to minimize the objective function while satisfying constraints. The genetic algorithm provides global search capabilities, utilizing selection, crossover, and mutation to improve population quality and enhance solution diversity. The flower pollination algorithm achieves a dynamic balance between global search and local optimization through Lévy flight and the golden sine mechanism, improving search accuracy and efficiency. The advantage of this approach is that the generated trajectory scheme not only satisfies physical constraints but also achieves an effective trade-off between multiple objectives such as running time, energy consumption, impact, and smoothness, significantly improving the efficiency, stability, and cleaning quality of the cleaning task.

[0093] S3: Detect and eliminate interference at the trajectory points on the determined cleaning trajectory.

[0094] In one possible implementation, S3 specifically includes sub-steps S301 to S307:

[0095] S301: Obtain the trajectory points and corresponding normal vectors on the cleaning trajectory.

[0096] S302: For each trajectory and its corresponding normal vector, construct the detection ray:

[0097] r(t) = v i +da i

[0098] Where r(t) represents the ray, d represents the distance along the normal vector, and v i Let a represent the i-th trajectory point. i Let represent the normal vector corresponding to the i-th trajectory point.

[0099] S303: Determine the intersection points of the detection ray and the point cloud mesh, and solve the equations.

[0100] v i+da i = (1-b1-b2)p0+b1p1+b2p2

[0101] Where p0, p1, and p2 represent the three vertices of the point cloud mesh, and b1 and b2 represent the coefficients to be solved.

[0102] S304: Using Cramer's rule, solve the equation for the intersection point and calculate d, b1, and b2:

[0103]

[0104] P1 = p1 - p0

[0105] P2 = p2 - p0

[0106] S = v i -p0

[0107] S1=a i ×P2

[0108] S2 = S × P1.

[0109] S305: Determine whether the solved b1 and b2 satisfy the following conditions: if yes, the ray intersects the point cloud mesh; otherwise, the ray does not intersect the point cloud mesh:

[0110] 0≤b1≤1

[0111] 0≤b2≤1

[0112] 0≤b1+b2≤1

[0113] S306: When a ray intersects with a point cloud mesh, the trajectory point is determined as an interference point.

[0114] S307: Adjust the cleaning trajectory according to the interference points to eliminate interference.

[0115] Specifically, the cleaning trajectory can be adjusted by interpolating from nearby non-interfering points.

[0116] In this invention, by constructing a detection ray, solving the intersection equations of the ray and the point cloud mesh, and judging the intersection conditions, the existence of interference is determined, and the trajectory points are dynamically adjusted when interference occurs. The advantages of this approach are: ensuring that the laser cleaning path does not collide with the surface of the component, improving the safety and reliability of the cleaning process; and simultaneously, avoiding cleaning deviations caused by interference through trajectory optimization, ensuring cleaning quality, and adapting to the cleaning needs of components with complex shapes.

[0117] S4: Construct the kinematic model of the robotic arm.

[0118] In one possible implementation, S4 specifically involves using Denavit-Hartenberg parameters to describe the geometric and kinematic relationships of each joint of the robotic arm and constructing a kinematic model of the robotic arm.

[0119] Denavit-Hartenberg (DH) parameters are a standardized method for modeling robotic arms, used to describe the geometric relationships and kinematic properties between the joints and links of a robotic arm. It constructs a homogeneous transformation matrix by defining the relative position and orientation of each joint and link, thereby enabling kinematic modeling and analysis of the robotic arm.

[0120] S5: Using the kinematic model of the robotic arm, the coordinates of the trajectory points on the cleaning trajectory are converted into the rotation angles of each joint of the robotic arm.

[0121] Specifically, based on the DH parameter table of the robotic arm, the coordinates of the trajectory points on the cleaning path can be initially converted into the rotation angles of each joint of the robotic arm. Further optimization can then be achieved using iterative solutions.

[0122] S6: With the goal of reducing the difference between the end effector of the robotic arm and the trajectory points on the cleaning path, the rotation angles of each joint of the robotic arm are optimized by using the Jacobian matrix in combination with iterative solution.

[0123] In one possible implementation, S6 specifically includes sub-steps S601 to S607:

[0124] S601: Get the current joint rotation angle value, convergence condition, and maximum number of iterations.

[0125] S602: Calculate the current pose matrix and Jacobian matrix.

[0126] The pose matrix represents the current position and orientation of the robotic arm's end effector, while the Jacobian matrix defines the relationship between joint velocities and end effector velocities.

[0127] Optionally, when the Jacobian matrix is ​​a singular matrix, the generalized inverse matrix is ​​used instead of the Jacobian matrix for calculation.

[0128] In this invention, when the Jacobian matrix is ​​a singular matrix, using a generalized inverse matrix can effectively avoid computational failures and ensure the continuity and stability of the inverse kinematics iteration process. The singularity of the Jacobian matrix typically occurs at certain specific locations (such as when the robotic arm reaches a motion boundary or certain poses), and direct inversion may cause the algorithm to break down. Introducing a generalized inverse matrix can provide a reasonable approximate solution near singular points, maintaining the updating of joint angles while preserving the accuracy of approximation to the target point. This method enhances the robustness of the algorithm, making it more adaptable to complex path planning.

[0129] S603: Calculate the difference motion matrix based on the current pose matrix:

[0130]

[0131] Where ΔT represents the difference motion matrix, T c T represents the pose matrix of the robotic arm's end effector. e Let I represent the pose matrix of the target point, and let I represent the identity matrix.

[0132] The elements of the difference motion matrix and the elements of the difference motion vector have the following relationship:

[0133]

[0134] Where dx represents the linear displacement component in the x-direction of Cartesian space, dy represents the linear displacement component in the y-direction of Cartesian space, dz represents the linear displacement component in the z-direction of Cartesian space, δx represents the angular displacement component in the x-direction of Cartesian space, δy represents the angular displacement component in the y-direction of Cartesian space, and δz represents the angular displacement component in the z-direction of Cartesian space.

[0135] S604: Calculate the difference motion vector based on the difference motion matrix:

[0136] D = [d x d y d z δ x δ y δ z ] T

[0137] Where D represents the difference motion vector and T represents the matrix transpose operation.

[0138] S605: Calculate the increment of joint angles based on the difference motion vector:

[0139] Δθ=J -1 D

[0140] Where Δθ represents the joint rotation angle increment, J represents the Jacobian matrix, -1 represents the inverse matrix operation, and D represents the difference motion vector calculated based on the difference motion matrix.

[0141] S606: Based on the increment of the joint angle, the joint rotation angle value is updated using an iterative solution method.

[0142] θ c (k+1)=θ c (k)+α(k)Δθ

[0143] Where, θ c (k+1) represents the joint rotation angle value at the (k+1)th iteration, θc (k) represents the joint rotation angle value at the k-th iteration, α(k) represents the adaptive learning rate at the k-th iteration, and Δθ represents the joint rotation angle increment.

[0144] The adaptive learning rate is as follows:

[0145]

[0146] Where α0 represents the initial learning rate, α K represents the maximum learning rate, and K represents the maximum number of iterations.

[0147] In this invention, an adaptive learning rate is introduced, enabling a dynamic balance between speed and accuracy during the optimization process. A larger initial learning rate helps to quickly approach the target; as the number of iterations increases, the learning rate decays exponentially, effectively reducing oscillations and finely adjusting joint angles to ensure convergence stability. This method avoids convergence failure caused by excessively large step sizes or slow convergence speed caused by excessively small step sizes, while adapting to the needs of different optimization stages, making the robotic arm joint optimization process more efficient and precise.

[0148] S607: Determine if the maximum number of iterations has been reached. If yes, output the rotation angle values ​​of each joint of the robotic arm. Otherwise, return to S601 and continue iterating.

[0149] In this invention, the rotation angles of each joint of the robotic arm are optimized by combining the Jacobian matrix with iterative solution, which can effectively reduce the difference between the end effector of the robotic arm and the target point on the cleaning trajectory. This method calculates the difference motion matrix and difference motion vector, and gradually approximates the target point using the joint angle increments. At the same time, an adaptive learning rate is introduced to control the iteration step size to improve the convergence speed and stability.

[0150] S7: Based on the mapping relationship between the trajectory points on the cleaning path and the rotation angles of each joint of the robotic arm, control the movement of the robotic arm and use lasers to automatically clean irregularly shaped parts.

[0151] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0152] In this invention, irregularly shaped parts are visually scanned to generate three-dimensional models of the parts. The goal is to reduce the running time, energy consumption, impact, and acceleration of the robotic arm. By combining flower pollination with a genetic optimization algorithm, an automatic cleaning trajectory is planned to improve the accuracy of the automatic cleaning trajectory. Then, according to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angles of each joint of the robotic arm, the movement of the robotic arm is controlled, and laser is used to automatically clean irregularly shaped parts, thereby improving the utilization rate of the robotic arm and improving the efficiency and accuracy of laser automatic cleaning.

[0153] Reference manual attached Figure 2 The diagram shows a schematic of the structure of an automatic laser cleaning trajectory planning system for irregularly shaped parts provided in an embodiment of the present invention.

[0154] The present invention also provides a laser automatic cleaning trajectory planning system 20 for irregularly shaped parts, comprising:

[0155] Processor 201;

[0156] The memory 202 stores computer-readable instructions, which, when executed by the processor 201, implement the laser automatic cleaning trajectory planning method for irregularly shaped parts as described in the method embodiment.

[0157] The laser automatic cleaning trajectory planning system 20 for irregularly shaped parts provided by the present invention can execute the above-mentioned laser automatic cleaning trajectory planning method for irregularly shaped parts and achieve the same or similar technical effects. To avoid repetition, the present invention will not elaborate further.

[0158] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0159] In this invention, the irregularly shaped parts are visually scanned to generate a three-dimensional model of the parts. With the goal of reducing the running time, energy consumption, impact, and acceleration of the robotic arm, an automatic cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm to improve the accuracy of the automatic cleaning trajectory. Then, according to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angles of each joint of the robotic arm, the movement of the robotic arm is controlled, and laser is used to automatically clean the irregularly shaped parts, thereby improving the utilization rate of the robotic arm and improving the efficiency and accuracy of laser automatic cleaning.

[0160] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

[0161] The following points need to be explained:

[0162] (1) The accompanying drawings of the embodiments of the present invention only involve the structures involved in the embodiments of the present invention. Other structures can refer to the general design.

[0163] (2) For clarity, the thickness of a region or area is enlarged or reduced in the drawings used to describe embodiments of the invention; that is, these drawings are not drawn to scale. It is understood that when an element such as a film, region, or substrate is referred to as being “above” or “below” another element, the element may be “directly” located “above” or “below” the other element or there may be intermediate elements.

[0164] (3) Where there is no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other to obtain new embodiments.

[0165] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for automatic laser cleaning trajectory planning of irregularly shaped parts, characterized in that, include: S1: Perform visual scanning on the irregularly shaped parts to generate a three-dimensional model of the parts; S2: Based on the point cloud coordinates on the surface of the three-dimensional model of the component, with the goal of reducing the running time, energy consumption, impact and acceleration of the robotic arm, a cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm. S3: Detect and eliminate interference at the trajectory points on the determined cleaning trajectory; S4: Construct the kinematic model of the robotic arm; S5: Using the kinematic model of the robotic arm, the coordinates of the trajectory points on the cleaning trajectory are converted into the rotation angles of each joint of the robotic arm; S6: With the goal of reducing the difference between the end effector of the robotic arm and the trajectory points on the cleaning trajectory, the rotation angles of each joint of the robotic arm are optimized by using the Jacobian matrix in combination with iterative solution. S7: According to the mapping relationship between the trajectory points on the cleaning trajectory and the rotation angle of each joint of the robotic arm, control the movement of the robotic arm and use laser to automatically clean irregular parts. Specifically, S6 includes: S601: Obtain the current joint rotation angle value, convergence condition, and maximum number of iterations; S602: Calculate the current pose matrix and Jacobian matrix; S603: Based on the current pose matrix, calculate the difference motion matrix: Where, Δ T Represents the difference motion matrix. T c This represents the pose matrix of the robotic arm's end effector. T e This represents the pose matrix of the target point. I Represents the identity matrix; The elements of the difference motion matrix and the elements of the difference motion vector have the following relationship: Where dx represents the linear displacement component in the x-direction of Cartesian space, dy represents the linear displacement component in the y-direction of Cartesian space, dz represents the linear displacement component in the z-direction of Cartesian space, δx represents the angular displacement component in the x-direction of Cartesian space, δy represents the angular displacement component in the y-direction of Cartesian space, and δz represents the angular displacement component in the z-direction of Cartesian space. S604: Based on the difference motion matrix, calculate the difference motion vector: in, D Represents the difference motion vector. T This represents the matrix transpose operation; S605: Based on the difference motion vector, calculate the increment of the joint angle: in, Indicates the increment of joint rotation angle. J This represents the Jacobian matrix, and -1 represents the inverse matrix operation. D This represents the difference motion vector calculated based on the difference motion matrix; S606: Based on the increment of the joint angle, update the joint rotation angle value using an iterative solution method: in, θ c ( k +1) indicates the first k The joint rotation angle value at +1 iteration. θ c ( k ) indicates the first k The joint rotation angle value at the next iteration. α ( k ) indicates the first k The adaptive learning rate at the next iteration, Δ θ Indicates the increment of joint rotation angle; The adaptive learning rate is as follows: in, α 0 represents the initial learning rate. α K This represents the maximum learning rate. K Indicates the maximum number of iterations; S607: Determine whether the convergence condition or the maximum number of iterations has been met; if yes, output the rotation angle values ​​of each joint of the current robotic arm; otherwise, return to S601 and continue iterating.

2. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 1, characterized in that, Specifically, S1 is: The irregularly shaped parts are visually scanned using a high-precision camera or reverse modeling equipment to generate a three-dimensional model of the parts.

3. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 1, characterized in that, S2 specifically includes: S201: To reduce the operating time, energy consumption, impact, and acceleration of the robotic arm, construct an optimization objective function; S202: Set constraints during the cleaning trajectory planning process; S203: Under the constraints of the above conditions, with the goal of minimizing the optimization objective function, a cleaning trajectory is planned by combining flower pollination with a genetic optimization algorithm.

4. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 3, characterized in that, The specific optimization objective function is as follows: Where min represents minimization. F This represents the objective function to be optimized. c 1 represents the robotic arm's runtime. c 2 represents the energy consumption of the robotic arm. c 3 indicates the impact on the robotic arm joint. c 4 represents the acceleration of the robotic arm joint. λ 1 represents the weighting coefficient of the robotic arm's running time. λ 2 represents the weighting factor for the robotic arm's energy consumption. λ 3 represents the weighting coefficient of the impact on the robotic arm joints. λ 4 represents the weighting coefficient of the acceleration of the robotic arm joints. ξ 1 represents the normalization factor for the robotic arm's runtime. ξ 2 represents the normalization factor for the energy consumption of the robotic arm. ξ 3 represents the normalization factor for the impact on the robotic arm joints. ξ 4 represents the normalization coefficient of the acceleration of the robotic arm joints.

5. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 4, characterized in that, The specific constraints are as follows: in, λ 1 represents the weighting coefficient of the robotic arm's running time. λ 2 represents the weighting factor for the robotic arm's energy consumption. λ 3 represents the weighting coefficient of the impact on the robotic arm joints. λ 4 represents the weighting coefficient of the acceleration of the robotic arm joints. ω ki Indicates the first i The joint in the first k Angular velocity on the segment cleaning trajectory, τ 1 represents the speed safety factor. W i Indicates the first i The maximum speed of each joint J i Indicates the first i Moment of inertia of each joint α ki Indicates the first i The joint in the first k Angular acceleration on the segment cleaning trajectory, τ 2 represents the safety factor for the driving torque. F i Indicates the first i The maximum driving torque of each joint.

6. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 1, characterized in that, S3 specifically includes: S301: Obtain the trajectory points and corresponding normal vectors on the cleaning trajectory; S302: For each trajectory and its corresponding normal vector, construct the detection ray: in, r ( t ) represents rays, t Represents the distance along the normal vector. v i Indicates the first i A trajectory point, a i Indicates the first i The normal vector corresponding to each trajectory point; S303: Determine the intersection point of the detection ray and the point cloud mesh, and solve the equation: in, p 0、 p 1. p 2 represents the three vertices of the point cloud mesh. b 1. b 2 represents the coefficient to be solved; S304: Solve the intersection point equation using Cramer's rule, and calculate... t , b 1 and b 2: ； S305: Determine the result obtained from the solution. b 1. b 2. Does the ray intersect the point cloud mesh if the following conditions are met? Otherwise, the ray does not intersect the point cloud mesh: S306: When a ray intersects with a point cloud mesh, the trajectory point is identified as an interference point; S307: Adjust the cleaning trajectory according to the interference point to eliminate the interference.

7. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 1, characterized in that, Specifically, S4 is: The Denavit-Hartenberg parameters are used to describe the geometric and kinematic relationships of the joints of the robotic arm, and a kinematic model of the robotic arm is constructed.

8. The laser automatic cleaning trajectory planning method for irregularly shaped parts according to claim 1, characterized in that, When the Jacobian matrix is ​​a singular matrix, the generalized inverse matrix is ​​used to replace the Jacobian matrix for calculation.

9. A laser automatic cleaning trajectory planning system for irregularly shaped parts, characterized in that, include: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement the laser automatic cleaning trajectory planning method for irregularly shaped parts as described in any one of claims 1 to 8.

Images