Hybrid robot path switching fairing method based on curvature and time optimization

Abstract

The invention provides a hybrid robot path switching fairing method based on curvature and time optimization. The hybrid robot path switching fairing method comprises the following steps that a machining path is divided into a plurality of element paths; parameterization is carried out on a switching section, a remaining section and a transition section in each primitive path section; a quantum particle swarm optimization algorithm is combined with a greedy algorithm to optimize the curvature of the switching section; the quantum particle swarm optimization is combined with a moving window planning method to optimize the machining time of each section; and interpolation is conducted on the geometric path to generate a motion program, then a servo motor driving instruction of a driving joint is obtained through inverse solution of the robot, and a machining task is completed. The method has the advantages that the corner curvature and the fairing deviation of the machining path can be optimized, the machining time can be optimized, the machining path is optimized in a segmented mode by combining the greedy algorithm and the moving window planning method, complex solution of a high-dimensional problem is avoided, and the machining efficiency is improved. And the machining quality and execution efficiency of the robot are improved.

Description

Technical Field

[0001] The present invention belongs to the field of robot processing, and in particular relates to a hybrid robot path transition smoothing method based on curvature and time optimization. Background Art

[0002] In recent years, the aerospace and energy industries have seen increasing demand for machining complex surface parts. Hybrid robots, with their advantages of high precision, high rigidity, large workspace, and reconfigurability, are widely used in the manufacture of core components. To achieve high precision and efficiency in robotic machining, in-depth research is needed on toolpath smoothing and motion trajectory optimization.

[0003] At present, most robot path smoothing algorithms can only achieve the continuous curvature change rate at the path transition point, that is, G 3 The tool path is continuous, but the acceleration of its trajectory is discontinuous, the tool path is not smooth, which will cause the robot to vibrate and greatly affect the machining performance. Therefore, generating high-order continuous tool paths for hybrid robots is still a problem that needs to be solved.

[0004] The curvature of the machining path directly affects the machining accuracy of the robot. A machining path with too large curvature will not only increase the difficulty of motion control and lead to a decrease in tracking accuracy, but also reduce the surface machining quality. Therefore, it is necessary to optimize the curvature of the machining path.

[0005] The time of trajectory planning can directly reflect the work efficiency of the robot and is also related to whether its work meets the time constraints and requirements of enterprise production. Optimizing time as an indicator of trajectory planning can improve the work efficiency of the robot and meet the needs of practical applications.

[0006] In addition, existing trajectory planning methods are difficult to efficiently optimize multiple trajectories, and the algorithm optimization process takes a long time, resulting in low robot work efficiency. Summary of the Invention

[0007] In view of this, the present invention aims to propose a hybrid robot path transition smoothing method based on curvature and time optimization to solve at least one problem existing in the above-mentioned prior art.

[0008] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0009] A hybrid robot path transition smoothing method based on curvature and time optimization includes the following steps: S1, dividing a processing path into a plurality of primitive paths;

[0010] S2. Parameterize the transfer segment, remaining segment, and transition segment in each primitive path segment using 5th-order NURBS curve, 5th-order B-spline, and 7th-order polynomial curve respectively;

[0011] S3. Set the optimization variables, objective functions, and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the greedy algorithm to optimize the curvature of the transition section;

[0012] S4. Set the optimization variables, objective functions and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the moving window planning method to optimize the processing time of each section;

[0013] S5. Finally, the geometric path is interpolated using the optimized motion law to generate a motion program, and then the servo motor drive instructions for the driving joint are obtained through the robot inverse solution to complete the processing task.

[0014] Furthermore, in step S1, the number of primitive path segments is consistent with the number of transfer endpoints.

[0015] Furthermore, in step S1 , the primitive path is in the form of remaining segment-transition segment-transfer segment-transition segment-remaining segment.

[0016] Furthermore, in step S2, in order to ensure that C 3 Continuously, using the differential characteristics of NURBS curves, a fifth-order NURBS curve with 7 control points is selected as the parametric synchronization curve of the transition segment, and the node vector is set to u = [0 0 0 0 0 0.5 1 1 1 1 1 1] Τ The control points are calculated using the corresponding formulas according to the three types of remaining segments (line-line, line-arc, arc-arc). The control points of line-line are calculated using the following formula:

[0017] Q3=P3

[0018] Q0=P3+1 s m s

[0019] Q1=Q0-c l,1 l s m s

[0020] Q2=Q1-c l,2 l s m s

[0021] Q4=Q5-c l,2 l e m e

[0022] Q5=Q6-c l,1 l e m e

[0023] Q6=P3+1 e m e ;

[0024] Among them, Q i (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight lines of the front and back segments; l s is the length of the remaining straight line of the previous segment; l e is the length of the remaining straight line in the latter section; m s P i Point to P i-1 Unit direction vector; m e P i-1 Point to P i The unit direction vector of m s opposite direction; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively;

[0025] The control points of a line-arc primitive path are calculated as follows:

[0026]

[0027] Among them, Q i (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight line and the remaining arc; l s is the length of the remaining straight line; m s P i Point to P i-1 Unit direction vector; m e is the unit direction vector of the tangent line at the starting point of the remaining arc; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively; θ e is the center angle of the arc to be transferred; c e is the center vector of the arc, r e Its radius; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Coincident Frenet frames;

[0028] The control points of an arc-arc primitive path are calculated as follows:

[0029]

[0030] Among them, Q i (i=0,…,6) represents the 7 control points required for the transition curve; P3 represents the intersection of the remaining arcs of the front and back segments; m s is the unit direction vector of the tangent line at the end point of the first remaining arc segment, m e is the unit direction vector of the tangent line at the starting point of the second remaining arc; θ s is the center angle of the first arc to be transferred, θ e is the center angle of the second arc to be transferred; c s is the center vector of the first arc, r s Its radius; c e is the center vector of the second arc, r e Its radius; n s , t s is the coordinate system K s The unit tangent vector, principal normal vector, K s Indicates that the arc P i-1 P i Fixed and the origin and point P i Coincident Frenet frame; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Overlapping Frenet frames.

[0031] Furthermore, in step S2, the remaining segment is the part that must be processed during the machining process and the path segment that the robot must strictly pass through. Therefore, in the remaining segment, to ensure that the robot passes through multiple interpolation points, a fifth-order B-spline is selected as the parametric synchronization curve, and the node vectors and control points are inversely calculated based on the interpolation points.

[0032] Furthermore, in step S2, to ensure the C between the transfer segment and the remaining segment 3 Continuity, a 7th-order polynomial curve that can constrain position, velocity, acceleration and acceleration is selected as the parameter synchronization curve of the transition segment. The initial and terminal constraints are set according to the transfer segment and the remaining segment.

[0033] Furthermore, in step S3 and step S4, the quantum particle swarm optimization algorithm includes the following steps: A1, initializing the particle population, generating a random population, determining the maximum shrinkage and expansion coefficient m and the minimum shrinkage and expansion coefficient n of the quantum particle swarm, and initializing the individual historical optimum and the group optimum;

[0034] A2. Calculate the fitness function value of the particles, compare the fitness function value of each initial particle, take the particle with the smallest function value as the global optimal particle, and select the individual optimal extreme value of each particle;

[0035] A3. Based on the superposition characteristics of the quantum particle group, update the current position of the particle in space according to the following formula;

[0036]

[0037] Among them, u ij (t+1) is a random number that obeys uniform distribution between [0,1]; r(t+1) is the number of symbols that obey uniform distribution; a(t) is the contraction and expansion coefficient of the quantum particle swarm algorithm, and its value decreases from m to n as the number of iterations increases; m best is the average individual optimal, and the specific expression is as follows:

[0038]

[0039] Where N is the total number of particles; M is the particle dimension; P i (t) is the optimal position of the i-th particle at time t, and the specific expression is as follows:

[0040] P ij (t+1)=f ij (t+1)P ij (t)+[1-f ij (t+1)]P gj (t),j=1,2,...,M;

[0041] Among them, P ij (t+1) is the optimal position of the individual at time t+1, P gj (t) is the global optimal position at time t; f ij (t+1) is a random number between [0,1] that follows a uniform distribution;

[0042] A4. Update the fitness value of the particle according to the updated particle position;

[0043] A5. Based on the corresponding fitness value update, update the individual extreme values ​​and global extreme values ​​in the population;

[0044] A6. Determine whether the algorithm has reached the maximum number of iterations. If so, the algorithm ends and outputs the solution set that meets the requirements. Otherwise, jump to step A3.

[0045] Furthermore, in step S3, in the single processing area, the optimization variables are set as: x = [ω3, ω4, ω5] Τ ;

[0046] where ωi ,i=3,4,5 are the middle three weights of the NURBS curve of the transition segment path; by adjusting the weight, the local shape of the curve is adjusted;

[0047] The objective function is:

[0048]

[0049] Where κ(u) is the curvature ε pos,i The position smoothing deviation refers to the Hausdorff distance between the inserted smoothing spline and the original path segment. i The position smoothing deviation at is expressed as:

[0050]

[0051] Where a * Indicates point A on the transfer path * The position vector of

[0052] The constraints are:

[0053]

[0054] Where, ε min and κ max are the minimum smoothing deviation and the maximum curvature, respectively.

[0055] Furthermore, in step S3, when the processing area is across different processing areas, the optimization variables set are consistent with those in the single processing area; the objective function is:

[0056]

[0057] The constraints are:

[0058]

[0059] Among them, ε min is the minimum fairing deviation, [x z ,y z ,z z ] is the coordinate of the obstacle.

[0060] Furthermore, in step S4, in the single processing area, the optimization variables set are: x = [t1, t2, t3, t4, t5];

[0061] Among them, t1, t2, t3, t4, and t5 are the five time periods corresponding to the primitive path;

[0062] The objective function is:

[0063]

[0064] Where n is the number of trajectory segments, t i is the time corresponding to the i-th moment;

[0065] The constraints are:

[0066]

[0067] in, is the contour error of the entire machining path, is the maximum value of the contour error set; v i is the velocity value of the i-th driving joint; a i is the acceleration value of the i-th driving joint; v max The maximum speed allowed for the driven joint; a max The maximum acceleration allowed for the driven joint.

[0068] Furthermore, in step S4, when the processing area is across different processing areas, the optimization variables and objective functions set are consistent with those of the single processing area; the constraints are:

[0069]

[0070] Furthermore, in step S3, the quantum particle swarm is combined with the greedy algorithm to optimize the curvature of the machining path, including the following steps:

[0071] B1. Initialize the control point weights of the current primitive path transition segment;

[0072] B2. Input the quantum particle swarm and obtain the current optimal weight according to the set objective function and constraints;

[0073] B3. Continue to optimize the weight of the next primitive path;

[0074] B4, output weights and corresponding curvature and smoothing deviations;

[0075] B5. Determine whether the optimization of the complete machining path is completed. If so, the algorithm ends and outputs the solution set of weight, curvature and smoothing deviation. Otherwise, jump to step B1.

[0076] Furthermore, in step S4, the quantum particle swarm optimization combined with the moving window planning method to optimize the processing trajectory time includes the following steps:

[0077] C1, initialize the first 5 segments of primitive path;

[0078] C2. Input the quantum particle swarm and obtain the current optimal time according to the set objective function and constraints;

[0079] C3, moving window, to ensure the processing trajectory C 3Continuous, set the fifth time segment of the previous primitive path to the window overlap part as the first time segment of the next primitive path;

[0080] C4, the 4 time segments of the primitive path after initialization;

[0081] C5. Input these 5 time periods into the quantum particle swarm to optimize the subsequent 4 time periods;

[0082] C6. Determine whether the optimization of the complete machining trajectory is completed. If so, the algorithm ends and outputs the solution set of the optimal time. Otherwise, jump to step C3.

[0083] Compared with the existing technology, the hybrid robot path smoothing method based on curvature and time optimization described in the present invention has the following advantages:

[0084] (1) The present invention discloses a hybrid robot path transition smoothing method based on curvature and time optimization. The present invention uses 5-order NURBS curves, 5-order B-splines, and 7-order polynomial curves to parameterize the transition segment path, the remaining segment path, and the transition segment path, respectively, to achieve C 3 continuous.

[0085] (2) The present invention describes a hybrid robot path transition smoothing method based on curvature and time optimization. The present invention uses an optimization algorithm to consider the different processing requirements of a single processing area and cross-processing areas, sets different optimization variables, objective functions, and constraints, and obtains the trajectory of optimal curvature and processing time.

[0086] (3) The present invention describes a hybrid robot path transition smoothing method based on curvature and time optimization. The present invention uses a quantum particle swarm algorithm combined with a greedy algorithm and a moving window planning method to achieve segmented optimization, thereby avoiding the complex solution of high-dimensional problems and further improving the optimization efficiency of the machining trajectory and the execution efficiency of the tool path. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] The accompanying drawings, which constitute part of the present invention, are provided to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are provided to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:

[0088] Figure 1 A schematic diagram of a process flow of the method described in an embodiment of the present invention;

[0089] Figure 2 for Figure 1 Schematic diagram of the path parameterization;

[0090] Figure 3 for Figure 1 Schematic diagram of efficient trajectory optimization method based on quantum particle swarm;

[0091] Figure 4 for Figure 1 Schematic diagram of CNC machining;

[0092] Figure 5 This is a schematic diagram of the quantum particle swarm algorithm process according to an embodiment of the present invention;

[0093] Figure 6 Schematic diagram of a processing path in a single processing area and across processing areas according to an embodiment of the present invention;

[0094] Figure 7 This is a schematic diagram of position smoothing deviation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0095] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0096] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0097] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0098] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0099] like Figures 1 to 7 As shown in FIG, a hybrid robot path transition smoothing method based on curvature and time optimization includes the following steps:

[0100] Step 1: Divide the processing path into several primitive paths;

[0101] In a preferred embodiment, if Figure 1 As shown, the number of primitive path segments is consistent with the number of transfer endpoints, and the format is remaining segment-transition segment-transition segment-transition segment-remaining segment.

[0102] Step 2: Use 5th-order NURBS curve, 5th-order B-spline and 7th-order polynomial curve to parameterize the transfer segment, residual segment and transition segment in each primitive path segment;

[0103] In a preferred embodiment, if Figure 1 As shown in Figure 2, the fifth-order B-spline is selected as the parameter synchronization curve for the remaining segment, and the node vectors and control points are inversely calculated based on the interpolation points.

[0104] In a preferred embodiment, if Figure 1 As shown in the figure, a fifth-order NURBS curve with 7 control points is selected as the parameter synchronization curve of the transition segment, and the node vector is set to u = [0 0 0 0 0 0.5 1 1 1 1 1 1] Τ The control points are calculated using the corresponding formulas according to the three types of remaining segments (line-line, line-arc, arc-arc):

[0105] The control points of the line-to-line are calculated as follows:

[0106] Q3=P3

[0107] Q0=P3+1 s m s

[0108] Q1=Q0-c l,1 l s m s

[0109] Q2=Q1-c l,2 l s m s

[0110] Q4=Q5-c l,2 l e m e

[0111] Q5=Q6-c l,1 l e m e

[0112] Q6=P3+1 e m e ;

[0113] Among them, Qi (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight lines of the front and back segments; l s is the length of the remaining straight line of the previous segment; l e is the length of the remaining straight line in the latter section; m s P i Point to P i-1 Unit direction vector; m e P i-1 Point to P i The unit direction vector of m s opposite direction; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively;

[0114] The control points of a line-arc primitive path are calculated as follows:

[0115]

[0116] Among them, Q i (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight line and the remaining arc; l s is the length of the remaining straight line; m s P i Point to P i-1 Unit direction vector; m e is the unit direction vector of the tangent line at the starting point of the remaining arc; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively; θ e is the center angle of the arc to be transferred; c e is the center vector of the arc, r e Its radius; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Coincident Frenet frames;

[0117] The control points of an arc-arc primitive path are calculated as follows:

[0118]

[0119] Among them, Q i (i=0,…,6) represents the 7 control points required for the transition curve; P3 represents the intersection of the remaining arcs of the front and back segments; m sis the unit direction vector of the tangent line at the end point of the first remaining arc segment, m e is the unit direction vector of the tangent line at the starting point of the second remaining arc; θ s is the center angle of the first arc to be transferred, θ e is the center angle of the second arc to be transferred; c s is the center vector of the first arc, r s Its radius; c e is the center vector of the second arc, r e Its radius; n s , t s is the coordinate system K s The unit tangent vector, principal normal vector, K s Indicates that the arc P i-1 P i Fixed and the origin and point P i Coincident Frenet frame; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Overlapping Frenet frames.

[0120] In a preferred embodiment, if Figure 1 As shown, a 7th-order polynomial curve is selected as the parameter synchronization curve in the transition section, and the initial and terminal constraints are set according to the transfer section and the remaining section.

[0121] Step 3: Set the optimization variables, objective functions and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the greedy algorithm to optimize the curvature of the transition section; for the schematic diagram of the single processing area and the cross-processing area, please refer to Figure 6 .

[0122] In a preferred embodiment, if Figure 1 As shown, in the single processing area, the optimization variables set are: x = [ω3, ω4, ω5] Τ ;

[0123] where ω i ,i=3,4,5 are the middle three weights of the NURBS curve of the transition path;

[0124] The objective function is:

[0125]

[0126] Where κ(u) is the curvature ε pos,iThe position smoothing deviation refers to the Hausdorff distance between the inserted smoothing spline and the original path segment. Please refer to Figure 4 and Figure 7 , point P i The position smoothing deviation at is expressed as: ε pos,i =max(ε1,ε2)

[0127]

[0128] Where a * Indicates point A on the transfer path * The position vector of

[0129] The constraints are:

[0130]

[0131] Where, ε min and κ max are the minimum smoothing deviation and the maximum curvature, respectively.

[0132] In a preferred embodiment, if Figure 1 As shown in the figure, when the processing area is across different processing areas, the optimization variables set are the same as those in the single processing area, and the objective function is:

[0133]

[0134] The constraints are:

[0135]

[0136] Among them, ε min is the minimum fairing deviation, [x z ,y z ,z z ] is the coordinate of the obstacle.

[0137] Also, please refer to Figure 5 After setting the optimization variables, objective function, and constraints, the quantum particle swarm is used to optimize the objective function. The specific steps are as follows:

[0138] A1: Initialize the particle population, generate a random population, determine the maximum contraction and expansion coefficient m and the minimum contraction and expansion coefficient n of the quantum particle population, and initialize the individual historical optimal and group optimal;

[0139] A2: Calculate the fitness function value of the particle, compare the fitness function value of each initial particle, select the particle with the smallest function value as the global optimal particle, and select the individual optimal extreme value of each particle; A3: Based on the superposition state characteristics of the quantum particle group, update the current position of the particle in space according to the following formula;

[0140]

[0141] Among them, u ij (t+1) is a random number that obeys uniform distribution between [0,1]; r(t+1) is the number of symbols that obey uniform distribution; a(t) is the contraction and expansion coefficient of the quantum particle swarm algorithm, and its value decreases from m to n as the number of iterations increases; m best is the average individual optimal, that is, the average value of all current individual optimal values. The specific expression is as follows:

[0142]

[0143] Among them, N is the total number of particles, that is, the number of individuals; M is the particle dimension; P i (t) is the optimal position of the i-th particle at time t, and the specific expression is as follows:

[0144] P ij (t+1)=f ij (t+1)P ij (t)+[1-f ij (t+1)]P gj (t),j=1,2,...,M;

[0145] Among them, P ij (t+1) is the optimal position of the individual at time t+1, P gj (t) is the global optimal position at time t; f ij (t+1) is a random number between [0,1] that follows a uniform distribution;

[0146] A4: Update the fitness value of the particle according to the updated particle position;

[0147] A5: Based on the corresponding fitness value update, update the individual extreme values ​​and global extreme values ​​in the population;

[0148] A6: Determine whether the algorithm has reached the maximum number of iterations. If so, the algorithm ends and outputs the solution set that meets the requirements. Otherwise, jump to step A3.

[0149] Optimizing the curvature of the machining path using a greedy algorithm includes the following steps:

[0150] B1: Initialize the control point weights of the current primitive path transition segment;

[0151] B2: Input quantum particle swarm and obtain the current optimal weight according to the set objective function and constraints; B3: Continue to optimize the weight of the next primitive path;

[0152] B4: Output weights and corresponding curvature and smoothing deviations;

[0153] B5: Determine whether the optimization of the complete machining path is completed. If so, the algorithm ends and outputs the solution set of weight, curvature and smoothing deviation. Otherwise, jump to step B1.

[0154] Step 4: Set the optimization variables, objective functions and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the moving window planning method to optimize the processing time of each section; for the schematic diagram of the single processing area and the cross-processing area, please refer to Figure 3 .

[0155] In a preferred embodiment, if Figure 1 As shown in the figure, in the single processing area, the optimization variables set are: x = [t1, t2, t3, t4, t5];

[0156] Among them, t1, t2, t3, t4, and t5 are the five time periods corresponding to the primitive path;

[0157] The objective function is:

[0158]

[0159] Where n is the number of trajectory segments, t i is the time corresponding to the i-th moment;

[0160] The constraints are:

[0161]

[0162] in, is the contour error of the entire machining path, is the maximum value of the contour error set; v i is the velocity value of the i-th driving joint; a i is the acceleration value of the i-th driving joint; v max The maximum speed allowed for the driven joint; a max The maximum acceleration value allowed for driving the joint;

[0163] In a preferred embodiment, if Figure 1 As shown in the figure, when the processing area is across different processing areas, the optimization variables and objective functions set are the same as those in the single processing area, and the constraints are:

[0164]

[0165] After setting the optimization variables, objective function, and constraints, use quantum particle swarm to optimize and solve the objective function. The specific steps are described in step three.

[0166] In a preferred embodiment, if Figure 1As shown in the figure, the optimization of the processing trajectory time by combining the moving window planning method includes the following steps:

[0167] C1: Initialize the first 5-segment primitive path;

[0168] C2: Input quantum particle swarm and obtain the current optimal time according to the set objective function and constraints;

[0169] C3: Move the window to ensure the processing trajectory C 3 Continuous, set the fifth time segment of the previous primitive path to the window overlap part as the first time segment of the next primitive path;

[0170] C4: 4 periods of time for the primitive path after initialization;

[0171] C5: Input these 5 time periods into the quantum particle swarm to optimize the subsequent 4 time periods;

[0172] C6: Determine whether the optimization of the complete machining trajectory is completed. If so, the algorithm ends and outputs the solution set of the optimal time. Otherwise, jump to step C3.

[0173] Step 5. Finally, the optimized motion law is used to interpolate the geometric path to generate a motion program, and then the servo motor drive instructions for the driving joint are obtained through the robot inverse solution to complete the processing task. This invention proposes a six-axis hybrid robot processing path smoothing method based on the optimal curvature and time, which solves the problem of smoothing the processing path of any processing path. 3 On a continuous basis, the curvature of the machining trajectory and the machining time are optimized. The segmentation strategy avoids the complex solution of high-dimensional problems, effectively improving the machining efficiency and quality.

[0174] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A hybrid robot path smoothing method based on curvature and time optimization, characterized by: The following steps are involved: S1, dividing the processing path into several primitive paths; S2. Parameterize the transfer segment, remaining segment, and transition segment in each primitive path segment using 5th-order NURBS curve, 5th-order B-spline, and 7th-order polynomial curve respectively; S3. Set the optimization variables, objective functions, and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the greedy algorithm to optimize the curvature of the transition section; S4. Set the optimization variables, objective functions and constraints for the single processing area and the cross-processing area respectively, and use the quantum particle swarm algorithm combined with the moving window planning method to optimize the processing time of each section; S5. Finally, the optimized motion law is used to interpolate the geometric path to generate a motion program, and then the servo motor drive instructions for the driving joint are obtained through the robot inverse solution to complete the processing task; In steps S3 and S4, the quantum particle swarm optimization algorithm includes the following steps: A1. Initialize the particle population, generate a random population, determine the maximum contraction and expansion coefficient m and the minimum contraction and expansion coefficient n of the quantum particle population, and initialize the individual historical optimal and group optimal; A2. Calculate the fitness function value of the particles, compare the fitness function value of each initial particle, take the particle with the smallest function value as the global optimal particle, and select the individual optimal extreme value of each particle; A3. Update the current position of the particles in space based on the superposition state characteristics of the quantum particle group; A4. Update the fitness value of the particle according to the updated particle position; A5. Based on the corresponding fitness value update, update the individual extreme values ​​and global extreme values ​​in the population; A6. Determine whether the algorithm has reached the maximum number of iterations. If so, the algorithm ends and outputs the solution set that meets the requirements. Otherwise, jump to step A3. In step S3, the quantum particle swarm optimization combined with the greedy algorithm optimizes the curvature of the machining path, including the following steps: B1. Initialize the control point weights of the current primitive path transition segment; B2. Input the quantum particle swarm and obtain the current optimal weight according to the set objective function and constraints; B3. Continue to optimize the weight of the next primitive path; B4, output weights and corresponding curvature and smoothing deviations; B5. Determine whether the optimization of the complete machining path is completed. If so, the algorithm ends and outputs the solution set of weight, curvature and smoothing deviation. Otherwise, jump to step B1. In step S4, the quantum particle swarm optimization combined with the moving window planning method to optimize the processing trajectory time includes the following steps: C1, initialize the first 5 segments of primitive path; C2. Input the quantum particle swarm and obtain the current optimal time according to the set objective function and constraints; C3, moving window, to ensure the processing trajectory C 3 Continuous, set the fifth time segment of the previous primitive path to the window overlap part as the first time segment of the next primitive path; C4, the 4 time segments of the primitive path after initialization; C5. Input these 5 time periods into the quantum particle swarm to optimize the subsequent 4 time periods; C6. Determine whether the optimization of the complete machining trajectory is completed. If so, the algorithm ends and outputs the solution set of the optimal time. Otherwise, jump to step C3.

2. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S1, the number of primitive path segments is consistent with the number of transfer endpoints.

3. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S1 , the primitive path is in the form of remaining segment-transition segment-transfer segment-transition segment-remaining segment.

4. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 3, characterized in that: In step S2, in order to satisfy C 3 Continuously, a fifth-order NURBS curve with 7 control points is selected as the parametric synchronization curve of the transition segment, and the node vector is set to u = [0 0 0 0 0 0.5 1 1 1 1 1 1] Τ ; The control points are calculated using the corresponding formulas according to the three types of remaining segments; The control points of a line-to-line are calculated as follows: Q3=P3 Q0=P3+l s m s Q1=Q0-c l,1 l s m s Q2=Q1-c l,2 l s m s Q4=Q5-c l,2 l e m e Q5=Q6-c l,1 l e m e Q6=P3+l e m e ; Among them, Q i (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight lines of the front and back segments; l s is the length of the remaining straight line of the previous segment; l e is the length of the remaining straight line in the latter section; m s P i Point to P i-1 Unit direction vector; m e P i-1 Point to P i The unit direction vector of m s opposite direction; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively; The control points of a line-arc primitive path are calculated as follows: Among them, Q i (i=0,…,6) represents the 7 control points required for the transfer curve; P3 represents the intersection of the remaining straight line and the remaining arc; l s is the length of the remaining straight line; m s P i Point to P i-1 Unit direction vector; m e is the unit direction vector of the tangent line at the starting point of the remaining arc; c l,1 、c l,2 is the straight line segment smoothing length ratio coefficient, which is 0.2 and 0.45 respectively; θ e is the center angle of the arc to be transferred; c e is the center vector of the arc, r e Its radius; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Coincident Frenet frames; The control points of an arc-arc primitive path are calculated as follows: Among them, Q i (i=0,…,6) represents the 7 control points required for the transition curve; P3 represents the intersection of the remaining arcs of the front and back segments; m s is the unit direction vector of the tangent line at the end point of the first remaining arc segment, m e is the unit direction vector of the tangent line at the starting point of the second remaining arc; θ s is the center angle of the first arc to be transferred, θ e is the center angle of the second arc to be transferred; c s is the center vector of the first arc, r s Its radius; c e is the center vector of the second arc, r e Its radius; n s , t s is the coordinate system K s The unit tangent vector, principal normal vector, K s Indicates that the arc P i-1 P i Fixed and the origin and point P i Coincident Frenet frame; n e , t e is the coordinate system K e The unit tangent vector, principal normal vector, K e Indicates that the arc P i P i+1 Fixed and the origin and point P i Overlapping Frenet frames.

5. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step A3, based on the superposition state characteristics of the quantum particle group, the current position of the particle in space is updated according to the following formula; Among them, u ij (t+1) is a random number that obeys uniform distribution between [0,1]; r(t+1) is the number of symbols that obey uniform distribution; a(t) is the contraction and expansion coefficient of the quantum particle swarm algorithm, and its value decreases from m to n as the number of iterations increases; m best is the average individual optimal, and the specific expression is as follows: Where N is the total number of particles; M is the particle dimension; P i (t) is the optimal position of the i-th particle at time t, and the specific expression is as follows: P ij (t+1)=f ij (t+1)P ij (t)+[1-f ij (t+1)]P gj (t),j=1,2,...,M; Among them, P ij (t+1) is the optimal position of the individual at time t+1, P gj (t) is the global optimal position at time t; f ij (t+1) is a random number between [0,1] that follows a uniform distribution.

6. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S3, in the single processing area, the optimization variables set are: x=[ω3,ω4,ω5] Τ ; where ω i ,i=3,4,5 are the middle three weights of the NURBS curve of the transition segment path; by adjusting the weight, the local shape of the curve is adjusted; The objective function is: Where κ(u) is the curvature ε pos,i The position smoothing deviation refers to the Hausdorff distance between the inserted smoothing spline and the original path segment. i The position smoothing deviation at is expressed as: Where a * Indicates point A on the transfer path * The position vector of The constraints are: Where, ε min and κ max are the minimum smoothing deviation and the maximum curvature, respectively.

7. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S3, when the processing area is across different processing areas, the optimization variables set are the same as those in the single processing area; the objective function is: The constraints are: Among them, ε min is the minimum fairing deviation, [x z ,y z ,z z ] is the coordinate of the obstacle.

8. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S4, in the single processing area, the optimization variables set are: x=[t1,t2,t3,t4,t5]; Among them, t1, t2, t3, t4, and t5 are the five time periods corresponding to the primitive path; The objective function is: Where n is the number of trajectory segments, t i is the time corresponding to the i-th moment; The constraints are: in, is the contour error of the entire machining path, is the maximum value of the contour error set; v i is the velocity value of the i-th driving joint; a i is the acceleration value of the i-th driving joint; v max The maximum speed allowed for the driven joint; a max The maximum acceleration allowed for the driven joint.

9. The method for path smoothing of a hybrid robot based on curvature and time optimization according to claim 1, characterized in that: In step S4, when the processing area is across different processing areas, the optimization variables and objective functions set are consistent with those of the single processing area; the constraints are:

Images