Mechanical arm and laser synchronous machining path planning method and device

Abstract

The invention relates to the technical field of automatic control, and discloses a path planning method and device for mechanical arm and laser synchronous machining, and the method comprises the steps: obtaining a workpiece three-dimensional model, a material and technological parameters, and extracting an irregular curved surface to generate a discrete point set; the tail end pose of the mechanical arm is calculated through inverse kinematics, and a preliminary joint angle and a track sequence are obtained. The optimal laser parameters are matched in combination with the parameters, a change curve is generated and synchronously mapped to a joint angle sequence, and a synchronous joint angle sequence is obtained through spline interpolation fine adjustment. And fitting a smooth track by using a B spline curve, and generating a motion track control instruction in combination with the synchronous joint angle. And the mechanical arm and the laser move according to instructions, the track compensation coefficient is dynamically adjusted when the environment changes, online compensation is conducted through the self-adaptive Kalman filtering algorithm, and a correction control instruction is obtained. According to the method, irregular curved surface machining details are extracted, the mechanical arm and the laser working sequence are aligned, the machining environment is sensed in real time, the adaptability of path planning is improved, and therefore efficient and accurate machining is achieved.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular, to a path planning method and device for synchronous processing of a robotic arm and a laser. Background Art

[0002] The technology of synchronous processing of a robotic arm and a laser is widely used in the field of industrial manufacturing. For example, in automobile manufacturing, the robotic arm cooperates with the laser for high-precision cutting and welding of parts, ensuring the precise assembly of automobile components; in the aerospace field, this technology is used to manufacture complex structural parts, such as the skin and frame of an aircraft wing, ensuring the strength and precision of the structural parts; in the electronics industry, the robotic arm and the laser work together to achieve cutting and welding of small parts of precision electronic components, meeting the high-precision requirements.

[0003] Currently, the synchronous processing of a robotic arm and a laser mainly relies on an industrial control computer to coordinate the movements of the robotic arm and the laser. For example, through the industrial control computer, according to the preset processing trajectory data, the movement path of the robotic arm and the turning on and off of the laser are precisely controlled, so as to achieve the synchronous operation of the two and complete complex processing tasks.

[0004] However, the existing technology has obvious deficiencies in path planning. The movement trajectories of the robotic arm and the laser mostly rely on preset programs and lack real-time adaptability to the shapes of complex workpieces and changes in the dynamic processing environment. For example, when processing an irregular curved surface, the fixed path planning method may cause the robotic arm to be unable to accurately track the predetermined trajectory, thereby affecting the processing accuracy and making it difficult to guarantee the processing quality in high-precision processing scenarios. Summary of the Invention

[0005] The present invention provides a path planning method and device for synchronous processing of a robotic arm and a laser, which improve the adaptability of path planning by extracting the processing details of irregular curved surfaces, aligning the working sequences of the robotic arm and the laser, and real-time sensing of the processing environment, so as to perform efficient and precise processing.

[0006] In the first aspect, the present invention provides a path planning method for synchronous processing of a robotic arm and a laser, mainly including: Obtaining the three-dimensional model data of the workpiece, the workpiece material parameters, the processing process parameters, obtaining the displacement ranges and dynamic constraint conditions of each joint of the robotic arm, and obtaining the processing environment parameters; Extracting the irregular curved surface from the three-dimensional model data, and generating a high-density discrete point set by using a point cloud sampling algorithm; According to the high-density discrete point set, using an inverse kinematics algorithm, calculating the target pose of the end effector of the robotic arm at each discrete point, and combining the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence; According to the workpiece material parameters and the machining process parameters, match the optimal laser working parameters from a pre-established laser processing parameter database, and generate a laser parameter variation curve that matches the preliminary trajectory sequence; Synchronously map the laser parameter variation curve and the preliminary joint angle sequence along the time axis, and use the spline interpolation algorithm to finely adjust the preliminary joint angle sequence to obtain a synchronized joint angle sequence; Adopt the B-spline curve fitting algorithm to smooth the preliminary trajectory sequence, and combine it with the synchronized joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robotic arm; The robotic arm moves according to the motion trajectory control instruction. When the machining environment parameters change, dynamically adjust the preset trajectory compensation coefficient, and perform online compensation on the trajectory instruction through the adaptive Kalman filter algorithm to obtain a corrected motion control instruction.

[0007] In an alternative implementation, to extract the irregular surface in the three-dimensional model data and generate a high-density discrete point set using the point cloud sampling algorithm, it includes: For the irregular surface, use the uniform point cloud sampling algorithm to generate an initial discrete point set; Judge whether the point set density of the initial discrete point set is lower than the preset density threshold. If so, use the local encryption sampling algorithm to process the initial discrete point set. If not, do not perform any processing to obtain enhanced discrete point set data; According to the enhanced discrete point set data, use the random forest algorithm to classify the irregular surface to obtain different feature regions, and obtain the geometric property distribution of the different feature regions; According to the geometric property distribution, determine the direction consistency data of the surface of the irregular surface through normal estimation; According to the direction consistency data, for the enhanced discrete point set, use the k-nearest neighbor algorithm to smooth the point set spacing to obtain a high-density discrete point set.

[0008] In an alternative implementation, according to the high-density discrete point set, use the inverse kinematics algorithm to calculate the target pose of the end effector of the robotic arm at each discrete point, and combine the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence, including: Obtain the discrete point set data, use the inverse kinematics algorithm to determine the target pose of the end effector of the robotic arm at each discrete point, and obtain the original joint angle sequence for the joint displacement range limit and dynamic constraint conditions; Group the original joint angle sequences using the k - means clustering algorithm, and obtain the classified angle sequences based on the matching degree between each group of the original joint angle sequences and the dynamic constraint conditions; Adjust the angle values in the classified angle sequences that exceed the joint displacement range limit to the maximum value of the joint displacement range, and perform smoothing processing using the interpolation algorithm to obtain the optimized angle sequences; Determine the direction consistency data of the end - effector of the robotic arm at each discrete point of the optimized angle sequences through normal estimation, and the robotic arm combines the direction consistency data to obtain the motion trajectory parameters; For the dynamic constraint conditions, use the support vector machine algorithm to perform anomaly detection on the motion trajectory parameters. If abnormal trajectory points are detected, adjust the optimized angle sequences through local resampling to obtain the preliminary joint angle sequences and the preliminary trajectory sequences.

[0009] In an alternative embodiment, the matching of the optimal laser working parameters from the pre - established laser processing parameter database according to the workpiece material parameters and the processing technology parameters to generate a laser parameter change curve matching the preliminary trajectory sequence includes: Use the k - nearest neighbor algorithm to retrieve the laser power, scanning speed, and focused spot parameters that match the workpiece material parameters and the processing technology parameters from the pre - established laser processing parameter database to obtain a preliminary laser parameter set; According to the preliminary laser parameter set, use the interpolation algorithm to process the preliminary trajectory sequence to generate laser power and scanning speed change sequences corresponding to the trajectory points, and obtain the initial parameter change curve; Use the support vector machine algorithm to detect the abnormal parameter points in the initial parameter change curve that do not conform to the processing technology parameters, and adjust the parameter values through local resampling to obtain the corrected parameter change curve; Map the processing technology parameter sequence to the discrete point set of the motion trajectory through grid division for the corrected parameter change curve to obtain a spatially optimized data set; Use the linear regression algorithm to perform smoothing processing on the spatially optimized data set to obtain the final laser parameter change curve.

[0010] In an alternative embodiment, the synchronization mapping of the laser parameter change curve and the preliminary joint angle sequence along the time axis, and the fine - tuning of the preliminary joint angle sequence using the spline interpolation algorithm includes: Perform smoothing processing on the joint angle sequence using the cubic spline interpolation algorithm to obtain a smoothed joint angle sequence; Map the smooth joint angle sequence to a motion speed sequence through time-axis alignment, and map the laser parameter change curve to a laser power change sequence; Use the Kalman filter algorithm to perform noise filtering on the motion speed sequence to obtain a smooth speed sequence; Use the dynamic time warping algorithm to analyze the time deviation between the smooth speed sequence and the laser power change sequence, and calibrate to obtain a calibrated synchronization sequence; Judge the direction consistency between the speed sequence and the power change in the calibrated synchronization sequence through normal estimation to obtain a synchronized joint angle sequence.

[0011] In an optional implementation manner, the robotic arm moves according to the motion trajectory control instruction. When the processing environment parameters change, dynamically adjust the preset trajectory compensation coefficient, and perform online compensation on the trajectory instruction through the adaptive Kalman filter algorithm to obtain a corrected motion control instruction, including: Monitor the change of the processing environment parameters in real time, and calculate the preliminary trajectory instruction adjustment amount using the preset initial compensation coefficient to obtain the first trajectory adjustment sequence; Dynamically adjust the compensation coefficient according to the time distribution characteristics of the first trajectory adjustment sequence, and combine the change trend of the processing environment parameters. Use the adaptive Kalman filter algorithm to smooth and correct the first trajectory adjustment sequence to obtain the second trajectory adjustment sequence; Suppress high-frequency noise through a filtering algorithm according to the trajectory instruction distribution of the second trajectory adjustment sequence to obtain a third trajectory adjustment sequence; For the control sequence continuity of the third trajectory adjustment sequence, perform secondary correction on the deviation of the trajectory instruction through online compensation to obtain a corrected motion control instruction.

[0012] In an optional implementation manner, it further includes: using high-precision encoders at the joints of the robotic arm and power sensors on the laser processing head to collect the actual rotation angles of each joint and the laser output parameters in real time, and comparing them with the corrected motion control instruction. If the deviation exceeds the preset threshold, trigger closed-loop feedback control to generate a feedback control instruction to adjust the motion trajectory of the robotic arm and the laser parameters in real time, specifically including: Obtain the actual rotation angles of each joint and the laser output parameters collected in real time through high-precision encoders and power sensors to obtain the original data sequence; Compare the original data sequence with the corrected motion control instruction, and use a deviation calculation method to determine the deviation value sequence; Judge whether the deviation value sequence exceeds the preset deviation threshold. If so, start the closed-loop feedback mechanism. If not, do not perform any processing and generate a preliminary feedback control sequence; For the preliminary feedback control sequence, a low-pass filtering algorithm is used to suppress high-frequency interference and obtain a smoothed control sequence; Combined with the changing trend of the actual rotation angle, the smoothed control sequence is adjusted through online compensation to obtain an optimized control sequence; Regarding the matching situation between the optimized control sequence and the laser output parameters, a proportional regulation method is used to correct the optimized control sequence to obtain a feedback control instruction.

[0013] The present invention provides a path planning device for synchronous machining of a robotic arm and a laser, mainly including: A data acquisition module, configured to acquire three-dimensional model data of a workpiece, workpiece material parameters, machining process parameters, acquire the displacement ranges and dynamic constraint conditions of each joint of the robotic arm, and acquire machining environment parameters; A point cloud generation module, configured to extract an irregular surface from the three-dimensional model data and generate a high-density discrete point set by using a point cloud sampling algorithm; A preliminary trajectory generation module, configured to calculate the target pose of the end effector of the robotic arm at each discrete point according to the high-density discrete point set by using an inverse kinematics algorithm, and combine the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence; A laser parameter matching module, configured to match optimal laser working parameters from a pre-established laser processing parameter database according to the workpiece material parameters and the machining process parameters, and generate a laser parameter change curve matching the preliminary trajectory sequence; A synchronous mapping module, configured to synchronously map the laser parameter change curve and the preliminary joint angle sequence along the time axis, and finely adjust the preliminary joint angle sequence by using a spline interpolation algorithm to obtain a synchronous joint angle sequence; A trajectory smoothing module, configured to smooth the preliminary trajectory sequence by using a B-spline curve fitting algorithm, and combine the synchronous joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robotic arm; A dynamic adjustment module, configured to enable the robotic arm to move according to the motion trajectory control instruction, dynamically adjust a preset trajectory compensation coefficient when the machining environment parameters change, and perform online compensation on the trajectory instruction through an adaptive Kalman filtering algorithm to obtain a corrected motion control instruction.

[0014] In a third aspect, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a path planning method for synchronous machining of a robotic arm and a laser as described in any one of the above.

[0015] Fourthly, the present invention also provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the path planning method for synchronized processing of a robotic arm and a laser as described in any one of the above.

[0016] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: The present invention discloses a precision machining method based on a robotic arm and a laser, including: Aiming at the machining problem of irregular curved surface workpieces, the present invention first obtains a high-density discrete point set through point cloud sampling, and calculates the robotic arm motion trajectory in combination with the inverse kinematics algorithm. Then, the optimal laser parameters are matched according to the workpiece characteristics, and they are synchronously mapped with the joint angle sequence; To improve the machining accuracy, the present invention uses the B-spline curve algorithm to optimize the trajectory and introduces adaptive Kalman filtering for real-time compensation, so as to realize the smoothness of the trajectory during the machining process and compensate for the possible errors during the machining process, realizing high-precision machining; In addition, the present invention also establishes a closed-loop feedback control mechanism, and dynamically adjusts the motion trajectory and laser parameters through real-time monitoring and comparison. This innovative method significantly improves the machining accuracy and efficiency of complex curved surface workpieces and provides a new technical solution for the precision manufacturing field. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of a path planning method for synchronized processing of a robotic arm and a laser according to the present invention.

[0018] Figure 2 It is a schematic structural diagram of a path planning device for synchronized processing of a robotic arm and a laser according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] To further understand the content of the present invention, the present invention will be described in detail in combination with the drawings and embodiments. The following further describes the present application in detail with reference to the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention and are not a limitation to the invention. In addition, it should be noted that only the parts related to the invention are shown in the drawings for the convenience of description.

[0020] Refer to Figure 1 , a path planning method for synchronized processing of a robotic arm and a laser in this embodiment may specifically include the following steps: Step S101, obtain the three-dimensional model data of the workpiece, the workpiece material parameters, the machining process parameters, obtain the displacement range and dynamic constraints of each joint of the robotic arm, and obtain the machining environment parameters; Step S102: Extract the irregular surface from the three-dimensional model data, and generate a high-density discrete point set using a point cloud sampling algorithm; Step S103: According to the high-density discrete point set, use the inverse kinematics algorithm to calculate the target pose of the end effector of the robotic arm at each discrete point, and combine the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence; Step S104: According to the workpiece material parameters and the processing process parameters, match the optimal laser working parameters from the pre-established laser processing parameter database, and generate a laser parameter change curve that matches the preliminary trajectory sequence; Step S105: Synchronously map the laser parameter change curve and the preliminary joint angle sequence along the time axis, and use the spline interpolation algorithm to fine-tune the preliminary joint angle sequence to obtain a synchronous joint angle sequence; Step S106: Use the B-spline curve fitting algorithm to smooth the preliminary trajectory sequence, and combine it with the synchronous joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robotic arm; Step S107: The robotic arm moves according to the motion trajectory control instruction. When the processing environment parameters change, dynamically adjust the preset trajectory compensation coefficient, and perform online compensation on the trajectory instruction through the adaptive Kalman filtering algorithm to obtain a corrected motion control instruction.

[0021] In step S101, obtain the three-dimensional model data of the workpiece, the workpiece material parameters, the processing process parameters, obtain the displacement range and dynamic constraint conditions of each joint of the robotic arm, and obtain the processing environment parameters.

[0022] It should be noted that the three-dimensional model data is a geometric shape description of the workpiece, including dimensions, shape, and surface information. It is the basis for path planning and determines the motion trajectories and processing ranges of the robotic arm and the laser. The workpiece material parameters include the physical properties (such as hardness, thermal conductivity, reflectivity) and chemical properties of the material. These parameters directly affect the laser processing effect, such as the penetration depth of the laser, the energy absorption rate, and the processing speed, etc. The processing process parameters involve specific requirements during the processing, such as processing accuracy, surface roughness, processing depth, etc. These parameters determine the process path and parameter selection of laser processing. The displacement range of each joint refers to the maximum and minimum angle ranges within which each joint of the robotic arm can move. This is a physical constraint condition for the motion planning of the robotic arm and determines the reachability and flexibility of the robotic arm. The dynamic constraint conditions include the load capacity, acceleration, speed limit, etc. of the robotic arm. These constraint conditions affect the motion smoothness and processing accuracy of the robotic arm and are factors that must be considered in path planning. The processing environment parameters involve external conditions during the processing, such as temperature, humidity, and light intensity. These parameters will affect the performance of the laser and the processing effect. For example, temperature changes may cause thermal expansion of the workpiece and affect the processing accuracy. At the same time, the environmental parameters may also affect the motion control of the robotic arm and the performance of the sensors.

[0023] In step S102, the irregular surfaces in the three-dimensional model data are extracted, and a high-density discrete point set is generated using a point cloud sampling algorithm, including: For the irregular surface, an initial discrete point set is generated using a uniform point cloud sampling algorithm; It is judged whether the point set density of the initial discrete point set is lower than a preset density threshold. If so, the initial discrete point set is processed using a local encryption sampling algorithm. If not, no processing is performed, and enhanced discrete point set data is obtained; Based on the enhanced discrete point set data, a random forest algorithm is used to classify the irregular surface to obtain different feature regions, and the geometric property distribution of the different feature regions is obtained; Based on the geometric property distribution, the direction consistency data of the surface of the irregular surface is determined through normal estimation; Based on the direction consistency data, for the enhanced discrete point set, the k-nearest neighbor algorithm is used to smooth the point set spacing to obtain a high-density discrete point set.

[0024] It should be noted that an irregular surface refers to a complex geometric shape that cannot be accurately described by a simple mathematical equation; the initial discrete point set refers to a discrete point set evenly distributed on the surface, which can reflect the geometric features and shape information of the surface; the preset density threshold is a key parameter that determines the accuracy of describing the irregular surface. If the preset density threshold is too high, it is difficult to accurately describe the irregular surface. If the preset density threshold is too low, excessive computing resources will be consumed to describe the irregular surface. Exemplarily, the preset density threshold is 1000 points per square meter; the enhanced discrete point set data is obtained by encrypting the sampling on the basis of the initial discrete point set to make the point set density meet the preset density threshold, and contains more points for describing the irregular surface; the geometric property distribution refers to the spatial distribution of curvature, concavity and convexity, and normal vector direction on the irregular surface; the direction consistency data is the normal vector direction information of the surface at each point, which reflects the matching degree of the normal vector directions of adjacent points and is used to determine the local surface direction of each point.

[0025] As a feasible implementation, calculate the normal vector and gradient of each point of the three-dimensional model of the workpiece. By deriving the relationship between the gradient of the surface function and the integral of the normal vector field, establish and solve the Poisson equation to obtain the surface function of the three-dimensional model. The value of this function at the points inside the model is 1, and the value at the points outside is 0. According to the Poisson equation, use the Poisson algorithm to automatically generate a closed triangular mesh surface model and obtain sample triangular points, and automatically segment the reconstructed triangular mesh surface using the segmentation threshold, and delete the unreasonable triangular patches on the non-entity, so as to obtain an irregular surface with smooth and correct edges.

[0026] As a feasible implementation, assume there is an irregular surface. Uniform point cloud sampling generates an initial discrete point set by setting a fixed spacing, such as taking a point every 5 centimeters.

[0027] It should be noted that the core of the local encryption sampling algorithm is to increase the sampling points in the area with low point set density to improve the data density and detail performance. Adopt the local encryption sampling algorithm to identify the low-density area and locally increase the sampling points.

[0028] For example, in the low-density area, supplement the point set by linear interpolation method to reduce the spacing from 5 centimeters to 2 centimeters, and generate enhanced discrete point set data. The new point set increases from 400 to 600, and the detail performance is richer.

[0029] Furthermore, according to the enhanced discrete point set data, the random forest algorithm can be used to classify the geometric property distribution of the irregular surface. The random forest can effectively identify the characteristic areas on the surface through the integration of multiple decision trees.

[0030] For example, in a mountain model, the curved surface can be divided into flat areas, slope areas, and peak areas. During implementation, local height differences, slopes, and other features of the point set can be extracted as inputs. The random forest determines the category of each point through a voting mechanism. Suppose among 50 points in a certain area, 40 are classified as slope areas, indicating that the geometric characteristics of this area tend to be inclined. Such classification results intuitively reflect the distribution law of the curved surface morphology and help understand its spatial characteristics.

[0031] In a feasible embodiment, normal estimation is used to confirm the consistency of the curved surface direction, which can help understand the local geometric characteristics of the curved surface. For each point, points within its local neighborhood are selected. The selection of the neighborhood can be based on a fixed radius or a fixed number of nearest neighbor points. Exemplarily, each point and the surrounding 5 points are used as the neighborhood of this point. The direction consistency data of each point is obtained through normal estimation of the neighborhood.

[0032] It should be noted that for areas with the same direction consistency data, k-nearest neighbor adjusts the point positions to make the spacing more uniform by analyzing the distances between each point in the area and its neighboring points.

[0033] For example, in a slope area, the average distance between a certain point and its 5 neighboring points is 3 cm, but the maximum deviation reaches 1 cm. Through k-nearest neighbor smoothing, the point positions can be adjusted to a deviation less than 0.5 cm. As a result, the point set may increase from 600 to 800, forming a high-density point cloud. The advantage of this processing is to eliminate noise points, improve the smoothness and consistency of the curved surface, and facilitate subsequent modeling or visualization.

[0034] In step S103, according to the high-density discrete point set, using the inverse kinematics algorithm, calculate the target pose of the end effector of the robotic arm at each discrete point, and combine the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence, including: Obtain the discrete point set data, use the inverse kinematics algorithm to determine the target pose of the end effector of the robotic arm at each discrete point, and for the joint displacement range limit and the dynamic constraint conditions, obtain the original joint angle sequence; Group the original joint angle sequence using the k-means clustering algorithm, and obtain the classified angle sequence according to the matching degree between each group of the original joint angle sequence and the dynamic constraint conditions; Adjust the angle values in the classified angle sequence that exceed the joint displacement range limit to the maximum value of the joint displacement range, and perform smoothing processing using the interpolation algorithm to obtain the optimized angle sequence; Determine the direction consistency data of the end effector of the robotic arm at each discrete point in the optimized angle sequence through normal estimation, and the robotic arm obtains the motion trajectory parameters by combining the direction consistency data; For the dynamic constraint conditions, the support vector machine algorithm is used to perform anomaly detection on the motion trajectory parameters. If an abnormal trajectory point is detected, the optimization angle sequence is adjusted through local resampling to obtain a preliminary joint angle sequence and a preliminary trajectory sequence.

[0035] It should be noted that the target pose refers to the position and orientation that the end effector of the robotic arm is expected to reach in three-dimensional space. Specifically, it includes two parts: position and orientation. Position: the coordinates of the end effector in space (usually represented by x, y, z), and orientation: the rotation state of the end effector, represented by a rotation matrix, Euler angles, or quaternions. The original joint angle sequence refers to the sequence of angle values of each joint of the robotic arm calculated through the inverse kinematics algorithm. These angle values describe the rotation angles of each joint that the robotic arm needs to pass through from the initial state to the target pose. The classified angle sequence refers to the joint angle sequence re-divided according to the dynamic constraint conditions and clustering results after clustering analysis of the original joint angle sequence. Specifically, the classified angle sequence divides the original data into multiple subsequences with similar characteristics, and each subsequence corresponds to a specific motion pattern or dynamic characteristic. The direction consistency data reflects the normal vector directions of the end effector at different positions and has a high degree of consistency between adjacent points, helping to ensure that the end effector maintains the correct pose during movement. The motion trajectory parameters refer to the detailed information describing the motion path of the end effector of the robotic arm in space, including parameters such as position, velocity, acceleration, and jerk. The preliminary joint angle sequence refers to the joint angle sequence obtained after anomaly detection, smoothing processing, and local resampling adjustment during the motion planning of the robotic arm. The preliminary trajectory sequence refers to the path sequence of the end effector of the robotic arm moving in space and the result after anomaly detection and smoothing adjustment.

[0036] As an implementation, inverse kinematics calculates the angles of each joint of the robotic arm based on the end position and orientation of the robotic arm.

[0037] For example, assume a three-degree-of-freedom robotic arm operating on a point cloud on an irregular surface. The discrete point set contains 500 points, and each point has three-dimensional coordinates. Inverse kinematics can calculate that when the end effector reaches a certain point (such as x = 10 cm, y = 5 cm, z = 2 cm), the joint angles may be 30 degrees, 45 degrees, and 60 degrees. Considering the joint displacement range limitations (such as joint 1 limited to 0 - 90 degrees) and dynamic constraints (such as the speed not exceeding 2 m / s), the original joint angle sequence will record the angle solutions for all points.

[0038] Specifically, the k-means clustering algorithm is used to group the original joint angle sequence. Exemplarily, in an angle sequence of 500 points, the angles of some points change drastically due to the steepness of the surface. K-means can divide the sequence into three groups: small angle changes in the gentle area, moderate angles in the transition area, and large angle fluctuations in the steep area. Each group has a different degree of matching with the dynamic constraints. For example, the speed requirement in the gentle area is low and the matching degree is high, while in the steep area, there may be overspeed and the matching degree is low. The classified angle sequence is thus formed, further reflecting the motion characteristics.

[0039] For example, if the angle of a point in the steep area is 95 degrees, exceeding the range of 0 - 90 degrees, it is adjusted to 90 degrees. The linear interpolation algorithm is used for smoothing processing to make the angles of adjacent points smoothly transition from 85 degrees to 90 degrees, avoiding sudden changes. The optimized angle sequence is thus generated, with improved smoothness.

[0040] For example, in the optimized angle sequence, the angle jump is reduced to within 2 degrees from 10 degrees. This processing ensures the continuous movement of the robotic arm, reduces vibrations, and improves the execution accuracy. The direction consistency data of each point in the optimized angle sequence is determined through normal estimation. As an implementation, based on the coordinates of 5 neighboring points in the discrete point set, the normal vector of each point is estimated.

[0041] For example, the trajectory parameter can be set such that the end moves 1 cm along the normal direction to keep the machining direction consistent. The advantage is to improve the geometric adaptability of the trajectory. For the dynamic constraints, the support vector machine algorithm is used to detect anomalies in the motion trajectory parameters.

[0042] For example, assuming that the speed of a point in the trajectory suddenly increases to 2.5 m / s, exceeding the 2 m / s constraint, the support vector machine marks it as an abnormal point through feature classification. The local resampling is used to adjust the optimized angle sequence, such as adding sampling points near this point and recalculating the angles to reduce the speed to 1.8 m / s. The preliminary joint angle sequence and the preliminary trajectory sequence are thus generated.

[0043] In step S104, according to the workpiece material parameters and the machining process parameters, the optimal laser working parameters are matched from a pre-established laser processing parameter database, and a laser parameter change curve matching the preliminary trajectory sequence is generated, including: The k-nearest neighbor algorithm is used to retrieve the laser power, scanning speed, and focused spot parameters matching the workpiece material parameters and the machining process parameters from the pre-established laser processing parameter database to obtain a preliminary laser parameter set; According to the preliminary laser parameter set, the interpolation algorithm is used to process the preliminary trajectory sequence to generate the laser power and scanning speed change sequences corresponding to the trajectory points, obtaining an initial parameter change curve; The support vector machine algorithm is used to detect abnormal parameter points in the initial parameter change curve that do not conform to the processing process parameters, and the parameter values are adjusted by local resampling to obtain a corrected parameter change curve; The corrected parameter change curve is used to map the processing process parameter sequence to the discrete point set of the motion trajectory through grid division to obtain a spatially optimized data set; The linear regression algorithm is used to smooth the spatially optimized data set to obtain the final laser parameter change curve.

[0044] It should be noted that the laser power refers to the optical energy output by the laser per unit time; the scanning speed refers to the speed at which the laser beam moves on the surface of the workpiece. The scanning speed directly affects the laser processing effect. A higher scanning speed can reduce heat input and improve processing efficiency, but too high a speed will lead to a decline in processing quality; the focused spot parameters include spot size, spot shape, and spot quality, which describe the characteristics of the laser beam after focusing. In the embodiments of the present invention, the spot size is in micrometers (μm), which determines the accuracy and power density of laser processing. A smaller spot size can increase the power density and is suitable for fine processing, but too small a spot will cause an increase in the heat affected zone.

[0045] Exemplarily, when using the k-nearest neighbor algorithm to retrieve matching parameters from the laser processing parameter database, a practical workpiece processing scenario is taken as an example. Assume that the workpiece material is aluminum alloy, and the processing process requires a surface roughness of less than 1 micrometer. The database stores thousands of historical processing records, including material parameters, thickness, laser power, and other information. The k-nearest neighbor algorithm calculates the Euclidean distance between the current workpiece parameters and the records in the database and finds 5 nearest neighbor data items as the preliminary laser parameter set.

[0046] It should be noted that the initial parameter change curve refers to the function curve of the laser power and scanning speed changing with time during the laser processing process. The initial parameter change curve is generated by interpolating the preliminary laser parameter set and the preliminary trajectory sequence, and is used to describe the dynamic changes of the laser power and scanning speed during the laser processing process. In a feasible implementation, when processing the preliminary trajectory sequence through the interpolation algorithm, the situation where the robotic arm processes along a curved path can be considered. Assume that the trajectory contains 100 discrete points, and the preliminary parameter set gives a power of 300W and a speed of 1m / s. The cubic spline interpolation algorithm is used to obtain the distance between points to generate a change sequence.

[0047] For example, the starting power is 300W, a certain point in the middle is adjusted to 320W due to an increase in curvature, and the end drops to 280W. The speed smoothly transitions from 1m / s to 1.2m / s. The initial parameter change curve is thus formed to adapt to the trajectory change.

[0048] It should be noted that the detection of abnormal parameter points usually adopts the Support Vector Machine (SVM) algorithm, especially the one-class support vector machine. By learning the distribution characteristics of normal data, abnormal parameter points that deviate from the normal processing parameter range can be identified.

[0049] Specifically, when the support vector machine detects abnormal parameter points, assume that the speed at a certain point in the initial curve suddenly increases to 1.5 m / s, exceeding the upper limit of 1.2 m / s required by the process. After detecting the abnormal parameter point, cubic spline interpolation is used to obtain the normal value to replace the abnormal parameter point in the initial parameter change curve. Cubic spline interpolation fits the initial parameter change curve through a piecewise polynomial function to generate a smooth parameter change curve. The specific steps are as follows: extract the abnormal parameter and the parameter values of its adjacent points; calculate the new parameter value of the abnormal point using the cubic spline interpolation formula; replace the original parameter value of the abnormal point to generate a corrected parameter change curve.

[0050] Preferably, when the grid division technology maps the processing process parameters, the trajectory discrete point set is divided into a 10x10 grid. Assume that the power change range within a certain grid is from 290 W to 310 W, and after mapping, a specific value is assigned to each point. For example, the points near the edge take 295 W. This spatially optimized data set reflects the correspondence between the trajectory and the parameters, facilitating subsequent processing.

[0051] In one embodiment, when the linear regression algorithm smooths the spatially optimized data set, assume that the power values of 5 points in a certain segment are 295 W, 300 W, 310 W, 305 W, and 298 W respectively. After regression, they may be adjusted to 297 W, 299 W, 302 W, 304 W, and 300 W. The final laser parameter change curve is smoother, avoiding parameter mutations and improving processing consistency.

[0052] For example, for an aluminum alloy workpiece, when the power is finely adjusted from 300 W to 302 W and the speed remains at 1.1 m / s, the surface roughness may be more uniform.

[0053] It can be understood that this smoothing process reduces the instantaneous impact of the laser on the material and prolongs the equipment life.

[0054] It should be noted that if the process requirement is to increase the processing depth, the power can be adjusted to 320 W through an extended scheme, and at the same time, the speed is reduced to 0.9 m / s to maintain the heat input balance. This diverse design meets different requirements.

[0055] In one embodiment, the corrected curve after abnormal detection may show that the power drops from 310 W to 305 W and the speed drops from 1.2 m / s to 1 m / s in a steep trajectory segment, matching the normal change. Multiple-sided verification shows that after parameter adjustment, the processing accuracy is improved and the scrap rate is reduced.

[0056] For example, the flat section has high consistency, and the steep section has strong adaptability. The final curve takes both efficiency and quality into account.

[0057] In step S105, synchronously map the laser parameter change curve and the preliminary joint angle sequence along the time axis, and use the spline interpolation algorithm to finely adjust the preliminary joint angle sequence to obtain a synchronized joint angle sequence, including: Perform smoothing processing on the joint angle sequence using the cubic spline interpolation algorithm to obtain a smoothed joint angle sequence; Map the smoothed joint angle sequence to a motion speed sequence through time axis alignment, and map the laser parameter change curve to a laser power change sequence; Perform noise filtering processing on the motion speed sequence using the Kalman filter algorithm to obtain a smoothed speed sequence; Use the dynamic time warping algorithm to analyze the time deviation between the smoothed speed sequence and the laser power change sequence, and calibrate to obtain a calibrated synchronization sequence; Judge the direction consistency between the speed sequence and the power change in the calibrated synchronization sequence through normal estimation to obtain a synchronized joint angle sequence.

[0058] Exemplarily, the cubic spline interpolation algorithm fits the data points between any two adjacent points of the joint angle sequence using a cubic polynomial to generate a smooth and continuous curve.

[0059] For example, assume that the robotic arm joint angle sequence contains 5 points: 10 degrees, 12 degrees, 15 degrees, 11 degrees, and 13 degrees. Direct connection may result in sharp turns. After using cubic spline interpolation, a smooth curve is generated, and the angle values may become 10 degrees, 11.5 degrees, 14 degrees, 12.5 degrees, and 13 degrees. This method ensures a natural transition of joint motion and avoids jitter during the operation of the robotic arm through piecewise polynomial fitting. In a feasible implementation, the angular velocity is obtained by dividing the angle difference between two adjacent points of the smoothed joint angle sequence by the time interval between the two adjacent points through time axis alignment, and the angular velocities of each point are combined in chronological order to obtain a motion speed sequence. Assume that the time interval is 0.1 second and the smoothed angle sequence is 10 degrees, 11.5 degrees, 14 degrees. The calculated speeds are 15 degrees / second and 25 degrees / second. Similarly, the laser parameter change curve is mapped to a laser power change sequence. For example, the initial power of 300W changes to 310W and 305W over time, which will not be elaborated here.

[0060] It should be noted that the Kalman filter is a recursive algorithm used to estimate the dynamic state of a system from noisy measurement data. When processing a sequence of motion speeds, the Kalman filter combines the system's prediction model and actual measurement data to continuously update the estimated value of the speed. By utilizing the system's prior knowledge (such as the motion model) and the statistical characteristics of the measurement data (such as the noise distribution), the optimal estimated value is calculated at each time step, thereby effectively filtering out noise and obtaining a smooth speed sequence. This method can provide accurate and smooth speed estimation results in real time.

[0061] Exemplarily, the original speed sequence might be 15 degrees per second, 25 degrees per second, 20 degrees per second, but due to sensor noise, it is doped with outliers such as 30 degrees per second. Through the prediction and update steps, the Kalman filter smoothly adjusts it to 15 degrees per second, 24 degrees per second, 21 degrees per second. This processing reduces noise interference and ensures the reliability of the speed data.

[0062] It should be noted that the dynamic time warping algorithm analyzes the time deviation between the smooth speed sequence and the laser power change sequence. First, it calculates the local distance between each point in the two sequences to form a distance matrix; using dynamic programming techniques, it searches for an optimal path in the distance matrix, and this path can minimize the cumulative distance between the two sequences; through the optimal path, the two sequences can be aligned on the time axis, thereby quantifying the time deviation between them; based on the optimal warping path, each point on the path can be converted into a time deviation sequence, representing the phase difference between the two sequences at different time points.

[0063] For example, the speed sequence is 15 degrees per second, 24 degrees per second, and the power sequence is 300W, 310W, but their sampling times may be misaligned. Through finding the best matching path, the dynamic time warping adjusts the time axis to make the speed of 24 degrees per second correspond precisely to the power of 310W, generating a calibrated synchronization sequence.

[0064] Preferably, the normal estimation technique determines the direction consistency between the speed sequence and the power change. Assuming the machining trajectory is a curve, the speed direction changes along the tangent, and the power needs to increase as the speed increases.

[0065] In one embodiment, when the speed increases from 15 degrees per second to 24 degrees per second and the power increases from 300W to 310W, the normal estimation confirms that their directions are consistent. If the power drops to 290W, there is a mismatch and the synchronous joint angle sequence needs to be adjusted. This consistency ensures the stability during the machining process.

[0066] Step S106, adopt the B-spline curve fitting algorithm to smooth the preliminary trajectory sequence, and combine it with the synchronous joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robotic arm.

[0067] The B-spline curve fitting algorithm fits discrete trajectory points into a continuous and smooth curve by means of piecewise polynomial interpolation. It has local adjustability and high-order continuity, can effectively reduce the impact and jitter in the movement of the robotic arm, and improve the machining accuracy and efficiency. The B-spline curve fitting performs uniform parameterization on the preliminary trajectory sequence and generates a knot vector. According to the preliminary trajectory sequence and the knot vector, the control points of the B-spline curve are calculated by the least squares method, and the final smooth trajectory is generated through the De Boor algorithm. Then, combined with the synchronous joint angle sequence, motion trajectory control instructions are generated to control the movement of the robotic arm and its end effector.

[0068] In step S107, the robotic arm moves according to the motion trajectory control instructions. When the machining environment parameters change, the preset trajectory compensation coefficient is dynamically adjusted, and the trajectory instructions are compensated online through the adaptive Kalman filtering algorithm to obtain corrected motion control instructions.

[0069] The changes in the machining environment parameters are monitored in real time. The preliminary trajectory instruction adjustment amount is calculated using the preset initial compensation coefficient to obtain the first trajectory adjustment sequence. According to the time distribution characteristics of the first trajectory adjustment sequence, the compensation coefficient is dynamically adjusted. Combining with the change trend of the machining environment parameters, the first trajectory adjustment sequence is smoothed and corrected using the adaptive Kalman filtering algorithm to obtain the second trajectory adjustment sequence. According to the trajectory instruction distribution of the second trajectory adjustment sequence, the high-frequency noise is suppressed through a filtering algorithm to obtain the third trajectory adjustment sequence. Regarding the continuity of the control sequence of the third trajectory adjustment sequence, the deviation of the trajectory instructions is corrected twice through online compensation to obtain corrected motion control instructions.

[0070] It should be noted that the trajectory instruction adjustment amount refers to the amount of adjustment made to the motion trajectory instructions of the robotic arm or the laser processing equipment according to the changes in the machining environment parameters monitored in real time. The trajectory instruction adjustment amount is used to compensate for the trajectory deviation caused by environmental changes (such as temperature and load changes) or system errors, thereby ensuring the accuracy and stability of the machining process.

[0071] It should be noted that the initial compensation coefficient is a key parameter that determines the trajectory instruction adjustment amount. If it is set too large, the robotic arm and the laser will be easily affected by the environment. If it is set too small, it will be difficult for the robotic arm and the laser to make adaptive adjustments according to the environment. Monitoring the changes in the machining environment parameters in real time is the basis for ensuring the accuracy of the trajectory instruction adjustment.

[0072] Exemplarily, in a machining scenario, machining environment parameters such as temperature, humidity, or machine tool vibration can directly affect the accuracy of the tool path. Suppose a CNC machine tool is machining an aluminum part, and the ambient temperature rises from 25 degrees Celsius to 30 degrees Celsius. Thermal expansion may cause the tool to shift by 0.02 millimeters. By using a preset initial compensation coefficient, such as 0.8, to calculate the initial adjustment amount of the trajectory instruction, a first trajectory adjustment sequence can be generated. This sequence records the adjustment values at each moment, such as adjusting the tool position by 0.01 millimeters per second, to cope with the impact of temperature changes. The advantage of this is that it can quickly respond to environmental changes and avoid the accumulation of machining errors. According to the time distribution characteristics of the first trajectory adjustment sequence, it is particularly important to dynamically adjust the compensation coefficient.

[0073] Specifically, if the sequence shows that the adjustment values exhibit periodic fluctuations in time, such as a large offset peak occurring every 10 seconds, it may be caused by machine tool vibration. At this time, the compensation coefficient can be dynamically adjusted from 0.8 to 1.2. Combining with the change trend of machining environment parameters, such as the increase in vibration frequency, an adaptive Kalman filter algorithm is used for smoothing correction.

[0074] It can be understood that the adaptive Kalman filter can effectively reduce the influence of random noise through prediction and update steps, and generate a second trajectory adjustment sequence.

[0075] For example, in the corrected sequence, the fluctuation range of the adjustment values is reduced from ±0.01 millimeters to ±0.005 millimeters, improving the stability of the trajectory. The resulting technical effect is a significant improvement in machining accuracy. In a feasible implementation, according to the trajectory instruction distribution of the second trajectory adjustment sequence, a filtering algorithm is needed to suppress high-frequency noise. Suppose there is noise with a frequency higher than 50 Hertz in the sequence, which may be caused by high-frequency interference from the motor. At this time, a low-pass filter can be used to filter out the high-frequency components, obtaining a third trajectory adjustment sequence.

[0076] For example, the original 50 tiny jitters per second are smoothed into a stable adjustment of 5 times per second, and the trajectory instruction distribution is more uniform. This processing can reduce the unnecessary oscillation of the machine tool, extend the equipment life, and ensure the surface finish of the machining. Regarding the control sequence continuity of the third trajectory adjustment sequence, online compensation is used for secondary correction of the trajectory instruction deviation.

[0077] Preferably, if the continuity analysis finds that the adjustment value of a certain segment of the trajectory instruction suddenly changes from 0.005 millimeters to 0.015 millimeters, it may be caused by sensor delay. Through online compensation, this mutation is detected in real time and smoothly transitioned, for example, adjusting the mutation value to 0.008 millimeters, generating the final corrected motion control instruction.

[0078] In one embodiment, this method can control the machining error within ±0.003 millimeters.

[0079] For example, when machining a precision gear, such smooth correction can ensure the consistency of the tooth profile and avoid assembly failures caused by deviations. Its beneficial effect lies in improving the reliability of control instructions and machining quality.

[0080] It should be noted that the implementation of each technical theme closely revolves around the core logic of machining environment and trajectory adjustment.

[0081] For example, the combined use of dynamic adjustment of compensation coefficients and adaptive Kalman filtering not only smooths the trajectory but also adapts to the unpredictability of environmental changes; the combination of filtering algorithms and online compensation forms a complete solution from noise suppression to continuity optimization. This multi-faceted supported design ensures the robustness and efficiency of trajectory instruction adjustment, ultimately achieving the goal of high-precision machining.

[0082] Using high-precision encoders at the joints of the robotic arm and power sensors on the laser processing head, the actual rotation angles of each joint and the laser output parameters are collected in real time and compared with the corrected motion control instructions. If the deviation exceeds the preset threshold, a closed-loop feedback control is triggered to generate a feedback control instruction for real-time adjustment of the robotic arm motion trajectory and laser parameters.

[0083] Obtain the actual rotation angles of each joint and the laser output parameters collected in real time by high-precision encoders and power sensors to obtain the original data sequence; compare the original data sequence with the corrected motion control instructions, and use a deviation calculation method to determine the deviation value sequence; judge whether the deviation value sequence exceeds the preset deviation threshold. If so, start the closed-loop feedback mechanism. If not, do nothing and generate a preliminary feedback control sequence; for the preliminary feedback control sequence, use a low-pass filtering algorithm to suppress high-frequency interference to obtain a smooth control sequence; combine the change trend of the actual rotation angle and adjust the smooth control sequence through online compensation to obtain an optimized control sequence; for the matching situation between the optimized control sequence and the laser output parameters, use a proportional adjustment method to correct the optimized control sequence to obtain a feedback control instruction; Exemplarily, obtaining the actual rotation angles of each joint and the laser output parameters collected in real time by high-precision encoders and power sensors to obtain the original data sequence is the basis of the entire control process.

[0084] It should be noted that the original data sequence is arranged according to the time sequence of collection for the actual rotation angle and the laser output parameters.

[0085] Exemplarily, in a robotic arm processing device, a high-precision encoder can collect 1000 joint rotation angle data per second. For example, the rotation angle of joint 1 changes from 10 degrees to 10.2 degrees, while the power sensor synchronously records that the laser output power slightly decreases from 500 watts to 498 watts. These data form the original data sequence, reflecting the real-time state of the device operation.

[0086] It can be understood that such high-frequency acquisition can capture subtle changes and provide a reliable basis for subsequent deviation analysis. Comparing the original data sequence with the corrected motion control and using a deviation calculation method to determine the deviation value sequence is the key to judging whether the system needs to be adjusted.

[0087] Specifically, assume that the preset instruction requires the joint rotation angle to remain at 10 degrees, while the actual value is 10.2 degrees, and the deviation value is 0.2 degrees; the laser power is preset at 500 watts and the actual value is 498 watts, with a deviation of 2 watts. Such a deviation value sequence intuitively reflects the situation where the system deviates from the expectation. In a feasible implementation, the deviation value sequence is arranged in time, such as being recorded every 0.1 seconds, forming a dynamic trend, which helps subsequent judgment. Judging whether the deviation value sequence exceeds the preset deviation threshold determines whether to start the closed-loop feedback mechanism.

[0088] Preferably, if the thresholds are set at 0.1 degree for rotation angle and 3 watts for power, among the above deviations of 0.2 degrees and 2 watts, the rotation angle exceeds the threshold while the power does not. At this time, the closed-loop feedback mechanism is started to generate a preliminary feedback control sequence.

[0089] For example, the sequence may indicate that joint 1 needs to be adjusted back by 0.15 degrees to approach the target value. Such a mechanism ensures that the system responds promptly to significant deviations and improves stability. For the preliminary feedback control sequence, a low-pass filtering algorithm is used to suppress high-frequency interference to obtain a smooth control sequence, which can effectively cope with the influence of noise.

[0090] In one embodiment, assume that there are 20 small jitters per second in the sequence, which are caused by motor operation interference. Low-pass filtering can smooth the adjustment frequency to 5 times per second.

[0091] For example, the original abrupt adjustment of 0.15 degrees becomes a gradual decrease of 0.05 degrees. Such smoothing reduces the unnecessary jitter of the robotic arm and helps to extend the device life. Combining with the changing trend of the actual rotation angle, the smooth control sequence is adjusted through online compensation to obtain an optimized control sequence, further improving the response accuracy.

[0092] Specifically, if the rotation angle continuously rises from 10.2 degrees to 10.3 degrees, indicating a trend deviation, online compensation can be dynamically adjusted according to the change rate.

[0093] For example, increasing the adjustment value from 0.05 degrees to 0.08 degrees forms an optimized sequence. This method adapts to the gradual change characteristics of the environment or load, ensuring the flexibility of control. Regarding the matching situation between the optimized control sequence and the laser output parameters, a proportional adjustment method is used for correction to obtain a feedback control instruction, which is an important step to achieve coordinated control.

[0094] For example, if the laser power of 498 watts does not fully match the rotation angle adjustment, the proportional adjustment can finely adjust the power to 499 watts while keeping the rotation angle adjustment at 0.08 degrees.

[0095] It should be noted that this matching correction ensures the consistency between the laser and the robotic arm movement during the processing, effectively improving the processing quality.

[0096] In one embodiment, it is assumed that when machining a metal sheet, the rotation angle deviation and power fluctuation may cause uneven cutting. Through the above process, from the original data acquisition to the generation of the feedback instruction, the system can quickly adjust the joint positions and the laser output.

[0097] For example, the rotation angle smoothly returns from 10.2 degrees to 10 degrees, and the power stabilizes at 500 watts. This multi-link coordinated control method not only improves the processing accuracy but also enhances the adaptability of the equipment, providing a reliable guarantee for high-precision tasks.

[0098] Referring to Figure 2 , the present invention provides a path planning device for synchronous machining of a robotic arm and a laser, mainly including: A data acquisition module, used to acquire the three-dimensional model data of the workpiece, the workpiece material parameters, the machining process parameters, acquire the displacement ranges and dynamic constraint conditions of each joint of the robotic arm, and acquire the machining environment parameters; A point cloud generation module, used to extract the irregular surface in the three-dimensional model data and generate a high-density discrete point set using a point cloud sampling algorithm; A preliminary trajectory generation module, used to calculate the target pose of the end effector of the robotic arm at each discrete point according to the high-density discrete point set using an inverse kinematics algorithm, and combine the displacement range and the dynamic constraint conditions to obtain a preliminary joint angle sequence and a preliminary trajectory sequence; A laser parameter matching module, used to match the optimal laser working parameters from a pre-established laser processing parameter database according to the workpiece material parameters and the machining process parameters, and generate a laser parameter change curve matching the preliminary trajectory sequence; A synchronous mapping module, used to synchronously map the laser parameter change curve and the preliminary joint angle sequence along the time axis, and finely adjust the preliminary joint angle sequence using a spline interpolation algorithm to obtain a synchronous joint angle sequence; A trajectory smoothing module, which is used to adopt a B-spline curve fitting algorithm to smooth the preliminary trajectory sequence, and combine the synchronous joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robotic arm; A dynamic adjustment module, which is used for the robotic arm to move according to the motion trajectory control instruction. When the processing environment parameters change, it dynamically adjusts the preset trajectory compensation coefficient, and performs online compensation on the trajectory instruction through an adaptive Kalman filtering algorithm to obtain a corrected motion control instruction.

[0099] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and supplements can be made, and these improvements and supplements should also be regarded as the protection scope of the present invention.

[0100] The embodiment of the present invention also provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a path planning program for synchronous processing of a robotic arm and a laser. When the processor executes the computer program, it implements the steps in the above-mentioned embodiments of the path planning method for synchronous processing of a robotic arm and a laser, such as Figure 1 The steps shown in S101. Alternatively, when the processor executes the computer program, it implements the functions of each module in the above-mentioned device embodiments, such as the dynamic adjustment module.

[0101] Exemplarily, the computer program can be divided into one or more modules. The one or more modules are stored in the memory and executed by the processor to complete the present invention. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the electronic device.

[0102] The electronic device can be a computing device such as a desktop computer, a notebook, a palm computer, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above components are only examples of the electronic device, and do not constitute a limitation on the electronic device. It may include more or fewer components than the above, or combine some components, or different components. For example, the electronic device may further include input / output devices, network access devices, a bus, etc.

[0103] The so-called processor may be a Central Processing Unit (CPU), or it may also 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. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the electronic device and connects various parts of the entire electronic device using various interfaces and circuits.

[0104] The memory can be used to store the computer programs and modules. The processor realizes various functions of the electronic device by running or executing the computer programs and modules stored in the memory and by calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, FlashCard, at least one magnetic disk storage device, flash device, or other volatile solid-state storage devices.

[0105] Among them, if the modules integrated in the electronic device are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0106] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between the modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement without creative efforts.

[0107] The above-described specific embodiments have further elaborated on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. It is particularly pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A path planning method for synchronous processing of a robot arm and a laser, characterized in that: The method comprises: Obtain the three-dimensional model data, material parameters, and processing parameters of the workpiece, the displacement range and dynamic constraints of each joint of the robot arm, and the processing environment parameters; Extracting irregular surfaces from the three-dimensional model data, and generating a high-density discrete point set using a point cloud sampling algorithm; According to the high-density discrete point set, an inverse kinematics algorithm is used to calculate the target posture of the end effector of the robot arm at each discrete point, and a preliminary joint angle sequence and a preliminary trajectory sequence are obtained by combining the displacement range and the dynamic constraint condition; According to the workpiece material parameters and the processing parameters, the optimal laser working parameters are matched from a pre-established laser processing parameter database to generate a laser parameter variation curve that matches the preliminary trajectory sequence; Synchronously mapping the laser parameter variation curve and the preliminary joint angle sequence according to the time axis, and fine-tuning the preliminary joint angle sequence using a spline interpolation algorithm to obtain a synchronized joint angle sequence; Using a B-spline curve fitting algorithm to smooth the preliminary trajectory sequence, and combining the synchronous joint angle sequence to obtain a motion trajectory control instruction to control the operation of the robot arm; The robot arm moves according to the motion trajectory control instruction. When the processing environment parameters change, the preset trajectory compensation coefficient is dynamically adjusted, and the trajectory instruction is compensated online through an adaptive Kalman filter algorithm to obtain a corrected motion control instruction.

2. The method according to claim 1, characterized in that The step of extracting the irregular surface in the three-dimensional model data and generating a high-density discrete point set using a point cloud sampling algorithm includes: For the irregular surface, a uniform point cloud sampling algorithm is used to generate an initial discrete point set; Determine whether the point set density of the initial discrete point set is lower than a preset density threshold, if so, use a local encrypted sampling algorithm to process the initial discrete point set, if not, do not perform any processing, and obtain enhanced discrete point set data; According to the enhanced discrete point set data, a random forest algorithm is used to classify the irregular surface to obtain different feature areas, and the geometric characteristic distribution of the different feature areas is obtained; Determining directional consistency data of the irregular curved surface by normal estimation according to the geometric characteristic distribution; According to the directional consistency data, for the enhanced discrete point set, a k-nearest neighbor algorithm is used to smooth the point set spacing to obtain a high-density discrete point set.

3. The method according to claim 1, characterized in that The method of calculating the target pose of the end effector of the manipulator at each discrete point by using an inverse kinematics algorithm based on the high-density discrete point set and obtaining a preliminary joint angle sequence and a preliminary trajectory sequence by combining the displacement range and the dynamic constraint condition includes: Obtain discrete point set data, use inverse kinematics algorithm to determine the target pose of the end effector of the robot arm at each discrete point, and obtain the original joint angle sequence based on the joint displacement range limit and dynamic constraints; The original joint angle sequences are grouped using a k-means clustering algorithm, and classified angle sequences are obtained according to the degree of matching between each group of the original joint angle sequences and the dynamic constraint conditions; The angle values ​​in the classified angle sequence that exceed the limit of the joint displacement range are adjusted to the maximum value of the joint displacement range, and smoothed by using an interpolation algorithm to obtain an optimized angle sequence; Determine the direction consistency data of the actuator at the end of the robotic arm at each discrete point of the optimized angle sequence by normal estimation, and obtain the motion trajectory parameters of the robotic arm in combination with the direction consistency data; According to the dynamic constraints, the support vector machine algorithm is used to perform abnormal detection on the motion trajectory parameters. If abnormal trajectory points are detected, the optimized angle sequence is adjusted by local resampling to obtain a preliminary joint angle sequence and a preliminary trajectory sequence.

4. The method according to claim 1, characterized in that The step of matching the optimal laser working parameters from a pre-established laser processing parameter database according to the workpiece material parameters and the processing process parameters to generate a laser parameter variation curve matching the preliminary trajectory sequence includes: Retrieving laser power, scanning speed and focus spot parameters matching the workpiece material parameters and the processing parameters from the pre-established laser processing parameter database using a k-nearest neighbor algorithm to obtain a preliminary laser parameter set; According to the preliminary laser parameter set, the preliminary trajectory sequence is processed by an interpolation algorithm to generate a laser power and scanning speed change sequence corresponding to the trajectory points, and an initial parameter change curve is obtained; Using a support vector machine algorithm to detect abnormal parameter points in the initial parameter change curve that are inconsistent with the processing parameters, and adjusting the parameter values ​​by local resampling to obtain a corrected parameter change curve; Mapping the processing parameter sequence to a discrete point set of the motion trajectory through grid division of the correction parameter variation curve to obtain a spatial optimization data set; The spatial optimization data set is smoothed using a linear regression algorithm to obtain a final laser parameter variation curve.

5. The method according to claim 1, characterized in that The step of synchronously mapping the laser parameter variation curve and the preliminary joint angle sequence according to the time axis, and fine-tuning the preliminary joint angle sequence using a spline interpolation algorithm to obtain a synchronous joint angle sequence includes: The joint angle sequence is smoothed by using a cubic spline interpolation algorithm to obtain a smoothed joint angle sequence; Mapping the smooth joint angle sequence into a motion speed sequence and mapping the laser parameter variation curve into a laser power variation sequence through time axis alignment; A Kalman filter algorithm is used to perform noise filtering on the motion speed sequence to obtain a smooth speed sequence; A dynamic time warping algorithm is used to analyze the time deviation between the smoothed velocity sequence and the laser power variation sequence, and a calibration synchronization sequence is obtained by calibration; The direction consistency of the speed sequence and the power change in the calibration synchronization sequence is determined by normal estimation to obtain a synchronization joint angle sequence.

6. The method according to claim 1, characterized in that The robot arm moves according to the motion trajectory control instruction, and when the processing environment parameters change, the preset trajectory compensation coefficient is dynamically adjusted, and the trajectory instruction is compensated online through an adaptive Kalman filter algorithm to obtain a corrected motion control instruction, including: Monitor the changes of the processing environment parameters in real time, calculate the preliminary trajectory instruction adjustment amount using a preset initial compensation coefficient, and obtain a first trajectory adjustment sequence; According to the time distribution characteristics of the first trajectory adjustment sequence, the compensation coefficient is dynamically adjusted, and in combination with the change trend of the processing environment parameters, the first trajectory adjustment sequence is smoothed and corrected by using an adaptive Kalman filter algorithm to obtain a second trajectory adjustment sequence; According to the trajectory instruction distribution of the second trajectory adjustment sequence, high-frequency noise is suppressed by a filtering algorithm to obtain a third trajectory adjustment sequence; In view of the control sequence continuity of the third trajectory adjustment sequence, the deviation of the trajectory instruction is corrected twice through online compensation to obtain a corrected motion control instruction.

7. The method according to claim 6, characterized in that Also includes: The high-precision encoders at the joints of the robot and the power sensors on the laser processing head are used to collect the actual rotation angles and laser output parameters of each joint in real time and compare them with the corrected motion control instructions. If the deviation exceeds the preset threshold, the closed-loop feedback control is triggered to generate feedback control instructions to adjust the robot motion trajectory and laser parameters in real time, including: The actual rotation angles of each joint and the laser output parameters collected in real time by high-precision encoders and power sensors are obtained to obtain the original data sequence; Comparing the original data sequence with the modified motion control instruction, using a deviation calculation method to determine a deviation value sequence; Determine whether the deviation value sequence exceeds a preset deviation threshold, if so, start a closed-loop feedback mechanism, if not, do nothing, and generate a preliminary feedback control sequence; For the preliminary feedback control sequence, a low-pass filtering algorithm is used to suppress high-frequency interference to obtain a smooth control sequence; In combination with the variation trend of the actual rotation angle, the smooth control sequence is adjusted through online compensation to obtain an optimized control sequence; According to the matching condition between the optimized control sequence and the laser output parameters, the optimized control sequence is corrected by using a proportional adjustment method to obtain a feedback control instruction.

8. A path planning device for synchronous processing of a robot arm and a laser, characterized in that: The device comprises: The data acquisition module is used to obtain the three-dimensional model data of the workpiece, the material parameters of the workpiece, the processing parameters, the displacement range and dynamic constraints of each joint of the robot arm, and the processing environment parameters; A point cloud generation module, used to extract irregular surfaces in the three-dimensional model data and generate a high-density discrete point set using a point cloud sampling algorithm; A preliminary trajectory generation module is used to calculate the target posture of the end effector of the manipulator at each discrete point according to the high-density discrete point set by using an inverse kinematics algorithm, and obtain a preliminary joint angle sequence and a preliminary trajectory sequence by combining the displacement range and the dynamic constraint condition; A laser parameter matching module, used to match the optimal laser working parameters from a pre-established laser processing parameter database according to the workpiece material parameters and the processing process parameters, and generate a laser parameter change curve matching the preliminary trajectory sequence; A synchronous mapping module, used for synchronously mapping the laser parameter variation curve and the preliminary joint angle sequence according to the time axis, and fine-tuning the preliminary joint angle sequence using a spline interpolation algorithm to obtain a synchronous joint angle sequence; A trajectory smoothing module, used for smoothing the preliminary trajectory sequence using a B-spline curve fitting algorithm, and obtaining a motion trajectory control instruction in combination with the synchronous joint angle sequence to control the operation of the robot arm; The dynamic adjustment module is used for the robot arm to move according to the motion trajectory control instruction. When the processing environment parameters change, the preset trajectory compensation coefficient is dynamically adjusted, and the trajectory instruction is compensated online through an adaptive Kalman filter algorithm to obtain a corrected motion control instruction.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a path planning method for synchronous processing of a robot arm and a laser is implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute a path planning method for synchronous processing of a robot arm and a laser as described in any one of claims 1 to 7.

Images