CN112666831A - Active control method for grinding and polishing contact force of robot - Google Patents
Active control method for grinding and polishing contact force of robot Download PDFInfo
- Publication number
- CN112666831A CN112666831A CN202011547920.7A CN202011547920A CN112666831A CN 112666831 A CN112666831 A CN 112666831A CN 202011547920 A CN202011547920 A CN 202011547920A CN 112666831 A CN112666831 A CN 112666831A
- Authority
- CN
- China
- Prior art keywords
- robot
- polishing
- grinding
- force
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention belongs to the field of robot machining, and particularly discloses an active control method for a grinding and polishing machining contact force of a robot. The method comprises the following steps: calculating expected polishing amount of the robot polishing process to generate a processing path, an expected pose and expected polishing force in the robot polishing process; acquiring a plurality of position points of the robot in the Z direction of a workpiece coordinate system in the process of moving along a processing path, and constructing a target function according to the position points and the impedance control key parameters; constructing constraint conditions of key parameters, and acquiring an optimal solution of the key parameters by adopting a weight-improved particle swarm algorithm; and updating and adjusting the pose and the grinding and polishing force of the grinding and polishing of the robot in real time, so that the force tracking error of the grinding and polishing of the robot along the processing path is minimum. The method effectively improves the tracking effect of impedance control on the expected force, improves the machining precision of the grinding and polishing of the robot, and ensures the surface consistency and good surface roughness of the machined workpiece.
Description
Technical Field
The invention belongs to the field of robot machining, and particularly relates to an active control method for a grinding and polishing machining contact force of a robot.
Background
The robot grinding and polishing technology has the advantages of high automation degree, strong adaptability, capability of engaging in high-risk operation and the like, and is widely applied to grinding and polishing operation of various complex parts. However, in the precision machining process of complex curved surface parts, the robot is limited by the requirements of low positioning precision of the robot, insufficient machining rigidity, complex operation environment, high dependence on an off-line planned path and the like, and the robot is often required to be equipped with force control in the grinding and polishing process. Therefore, in the grinding and polishing process of the robot, in order to meet the requirement of processing quality, the contact force is generally required to be accurately controlled, and the constant-force grinding of the workpiece is realized, so that the processing consistency and better surface roughness are ensured.
In order to solve the above problems, patent document CN110561237A discloses a robot belt grinding method and system combining active and passive force control, which combines an active force sensor and a kalman filtering method, and respectively controls the posture of the robot and the normal contact force of a contact wheel to realize precise and comprehensive control of the contact force, thereby improving the surface precision of the workpiece processing surface, but the real-time performance is still poor, and the processing hysteresis ratio is relatively serious.
Patent document CN111015661A discloses an active vibration control method for a flexible load of a robot, which takes a time lag of a pre-constructed input shaper as a time constraint, accelerates an original reference trajectory, and then shapes the accelerated trajectory with the input shaper to obtain a shaped trajectory as an output trajectory of the robot, thereby reducing the control cost of the robot and enhancing the stability of a control system.
Patent document CN110549326A discloses a method for adjusting pose of grinding and polishing of a robot based on multiple active compliant controllers, which includes installing multiple active compliant controllers at the end of the robot, calibrating a tool coordinate system, and adjusting the position of the tool coordinate system to make the displacement value of each grinding and polishing tool on the active compliant controller smaller than a set value, thereby improving the bonding effect between the grinding and polishing tool and the workpiece, and significantly improving the processing efficiency of the grinding and polishing of the robot.
Disclosure of Invention
Aiming at the defects or the improvement requirements of the prior art, the invention provides an active control method of the contact force of the grinding and polishing processing of the robot, wherein the active control method of the contact force of the grinding and polishing processing of the robot is correspondingly designed by combining the characteristics of the grinding and polishing processing of the robot and the process characteristics of the active control of the force, the position and the attitude of the robot are controlled by using an impedance control strategy and a six-dimensional force sensor, and the key parameters of a control system, such as M, D, K and the like, are optimized by using a weight-improved particle swarm algorithm, so that the force control algorithm of the robot system is optimized, the active control of an abrasive belt grinding and polishing system of the robot is realized, and the tracking effect of the impedance control on the expected force is effectively improved on the one hand; on the other hand, the processing precision of the grinding and polishing of the robot is effectively improved, and the surface consistency and good surface roughness of the processed workpiece are ensured.
In order to achieve the purpose, the invention provides an active control method for the contact force of grinding and polishing of a robot, which comprises the following steps:
s1, calculating expected polishing amount and processing path of the robot polishing processing according to the appearance of the workpiece to be processed and the three-dimensional model of the workpiece, and generating expected pose and expected polishing force in the robot polishing processing process according to the inverse kinematics model of the robot polishing processing;
s2, simplifying the robot into a mass-damping-spring system under the condition of no impedance control, acquiring a plurality of position points of the robot in the Z direction of a workpiece coordinate system in the process of moving along the processing path, and constructing a target function according to the plurality of position points and key parameters of impedance control in the grinding and polishing processing of the robot;
s3, constructing a constraint condition of the key parameter, and acquiring an optimal solution of the key parameter by adopting a weight-improved particle swarm algorithm according to the constraint condition and an objective function;
and S4, updating and adjusting the pose and the grinding and polishing force of the grinding and polishing robot in real time according to the optimal solution of the key parameters, the expected grinding and polishing amount and the expected pose of the grinding and polishing robot, the actual pose of the robot and the actual grinding and polishing force in the grinding and polishing process of the robot, so that the force tracking error of the grinding and polishing process of the robot along the processing path is minimum, and further the grinding and polishing process of the workpiece to be processed with the expected grinding and polishing force is realized.
Preferably, in step S1, the specific steps of calculating the expected polishing amount of the robot polishing process according to the shape of the workpiece to be processed and the three-dimensional model of the workpiece are as follows:
s11, calibrating the robot, the workpiece coordinate system and the force sensor for measuring the actual grinding and polishing force;
s12, performing off-line programming and trajectory planning on the workpiece through the three-dimensional model of the workpiece to generate a processing path in the grinding and polishing process of the robot;
s13, point cloud data of the appearance of the workpiece to be processed is obtained through measurement, and the point cloud data are matched with the three-dimensional model of the workpiece, so that the expected polishing amount of the robot for polishing is obtained.
Further preferably, in step S2, the calculation model of the objective function is:
wherein N is the total number of position points continuously collected along the processing path, k is a position point sampling sequence index, fd(k) F (k) is a feedback value of the normal force of the robot at the single position point, and the calculation formula of the feedback value is as follows:
where M is the mass in the impedance controlMatrix, D damping matrix in impedance control, KpIs an elastic matrix in impedance control, s (k),
Respectively displacement, velocity and acceleration of the robot in the normal direction of the workpiece, KeIs an elastic matrix in contact with the environment.
Further preferably, in step S3, the calculation model of the constraint function is:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, KpIs an elastic matrix in impedance control.
Preferably, in step S4, in the grinding and polishing process, the robot continuously updates the optimal solution of the key parameter to implement force control in the grinding and polishing process of the robot, so as to update and adjust the pose and the grinding and polishing force of the grinding and polishing process of the robot in real time, and the calculation model of the differential equation of the robot motion in the force control process is:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, and KpFor the elastic matrix in impedance control, x (t),
Respectively the displacement, the speed and the acceleration of the robot in the normal direction of the workpiece, wherein l is the initial distance between the workpiece and the grinding and polishing wheel, and v is the initial speed when the workpiece moves towards the grinding and polishing wheel in the opposite direction.
More preferably, step S3 specifically includes the following steps:
s31, collecting the original data of the key parameters at each position point, and carrying out initialization processing to obtain randomly generated initialization particles and the positions and speeds of the particles;
s32, evaluating the fitness of each particle according to the objective function, and storing the position and the fitness value of the particle in the individual extreme value p of the particlebestIn (1), all individuals are extremized by pbestThe position of the particle with the optimal adaptive value and the adaptive value are stored in the global extreme value gbestPerforming the following steps;
s33, updating the position and the speed of the particle according to the constraint condition;
s34, updating the inertia weight in the particle swarm algorithm;
s35, comparing the adaptive value of each particle with the best position of the particle, if the adaptive value does not meet the preset requirement, returning to the step S33, if the adaptive value meets the preset requirement, taking the current value as the best position of the particle, and comparing the individual extreme values p of all the current particlesbestAnd global extreme gbestAnd updating the current global extremum gbest；
S36, when the particle swarm algorithm reaches a stop condition, stopping searching and outputting the optimal solution of the key parameters; otherwise, the step S33 is returned to continue the search.
More preferably, in step S33, the position and velocity of the particle are updated using the following equations:
wherein, c1And c2Is the number of accelerations, r1、r2Two are in [0,1 ]]Acceleration weight coefficient, x, varying in betweeni,j(t) is the ith particle at tDisplacement value of time vi,j(t) is the velocity value of the ith particle at time t, w is the nonlinear dynamic inertia weight value, pi,jFor individual extrema of different particles, pg,jAnd d is the dimension of the search space.
More preferably, in step S34, the inertia weight is updated by using the following nonlinear dynamic inertia weight equation:
wherein f represents a real-time objective function value of the particle; f. ofavgAnd fminRespectively representing the average value and the minimum target value of all the current particles, w is a nonlinear dynamic inertia weight value, wminIs the minimum value of the inertial weight, wmaxIs the maximum value of the inertial weight.
More preferably, in step S34, the inertia weight is updated based on the particle objective function values, and the inertia weight is decreased when the particle objective function values are dispersed, and the inertia weight is increased when the particle objective function values are matched.
Generally, compared with the prior art, the above technical solution conceived by the present invention mainly has the following technical advantages:
1. the robot posture is controlled by combining an impedance control strategy with a six-dimensional force sensor, key parameters such as M, D, K and the like of a control system are optimized by using a weight-improved particle swarm algorithm, and then a force control algorithm of the robot system is optimized, so that the active control of the contact force of a robot abrasive belt grinding and polishing system is realized, and on one hand, the tracking effect of impedance control on expected force is effectively improved; on the other hand, the processing precision of the grinding and polishing of the robot is effectively improved, and the surface consistency and good surface roughness of the processed workpiece are ensured.
2. The method of the invention introduces an impedance control strategy for indirectly controlling the position and the contact force of the robot by adjusting mechanical impedance, realizes the active control of the robot by combining the force perception and the feedback of the six-dimensional force sensor, and can monitor the processing condition in real time in the grinding and polishing process of the robot.
3. According to the method, the weight-improved particle swarm optimization is introduced to optimize the impedance control key parameters of the robot, the force tracking effect of the impedance control strategy is improved, the normal grinding and polishing force of the robot is more stable in the constant-force grinding and polishing control process, the machining precision of the grinding and polishing machining of the robot is improved, and the surface consistency and good surface roughness of a machined workpiece are ensured.
Drawings
Fig. 1 is a process schematic diagram of an active control method for a grinding and polishing contact force of a robot according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a robot impedance control and environment contact process model in an active control method for a robot polishing contact force according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of the impedance control of the robot in the active control method of the contact force of the grinding and polishing of the robot according to the embodiment of the present invention;
fig. 4 is a flow chart of a particle swarm algorithm with improved weights according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
As shown in fig. 1, fig. 2, fig. 3 and fig. 4, an active control method for a contact force of a robot polishing process provided by an embodiment of the present invention includes the following steps:
firstly, performing off-line programming on a workpiece, namely a CAD model of a blade in RobotStudio software, planning a machining path, and performing simulation verification;
secondly, completing the calibration of key coordinate systems such as a scanner, a grinding and polishing wheel, a workpiece coordinate system and the like through field software and a robot demonstrator;
secondly, calibrating sensors such as the robot and the six-dimensional force sensor respectively;
secondly, acquiring the position points of the robot in the z direction of a tool coordinate system in the robot motion process under the condition of no impedance control, performing polynomial fitting on the acquired position points, and determining key parameters in a target function so as to determine the target function of the primary constraint condition;
then, M, D, Kp parameters of the robot system are optimized by using a weight-improved particle swarm optimization algorithm, an optimal solution of impedance force control parameters for workpiece grinding and polishing processing is obtained, and then a force control algorithm of the robot is optimized;
then, in the field processing process, the robot clamps the workpiece, the scanner is used for scanning to obtain complete point cloud data of the workpiece, point cloud matching is carried out to obtain the correction amount of the workpiece on the original CAD model, and the result is output to the robot;
and finally, the robot system monitors the grinding and polishing force change in real time through the force sensor, and calls a force control algorithm according to the feedback of the processing force to adjust the pose of the robot, so that the grinding and polishing with the expected force are realized.
The active control method comprises the following specific steps:
calculating expected polishing amount and a processing path of the robot polishing processing according to the appearance of the workpiece to be processed and the three-dimensional model of the workpiece, and generating expected pose and expected polishing force in the robot polishing processing process according to the inverse kinematics model of the robot polishing processing. Namely, offline programming and simulation are carried out on a blade model in RobotStaudio software, and calibration of sensors such as a robot, a workpiece coordinate system, a scanner, a six-dimensional force sensor and the like is completed on site. In the field processing process, the robot clamps the workpiece, the scanner is used for scanning to obtain complete point cloud data of the workpiece, point cloud matching is carried out to obtain the correction quantity of the workpiece on the original CAD model, namely the expected polishing quantity, and the result is output to the robot.
Specifically, a robot, a workpiece coordinate system and a force sensor for measuring actual grinding and polishing force are calibrated, a workpiece is subjected to off-line programming and track planning through a three-dimensional model of the workpiece to generate a processing path in the grinding and polishing processing process of the robot, point cloud data of the appearance of the workpiece to be processed is obtained through measurement, and the point cloud data is matched with the three-dimensional model of the workpiece to obtain expected grinding and polishing amount of the grinding and polishing processing of the robot. And calibrating sensors such as the robot, a workpiece coordinate system, a scanner and a six-dimensional force sensor on a robot grinding and polishing processing site to obtain the position conversion relation among the sensors such as the robot, the workpiece coordinate system, the scanner and the six-dimensional force sensor. Scanning the appearance of a workpiece to be processed by adopting a scanner so as to obtain complete point cloud data of the workpiece to be processed, matching coordinate value information of each point of a workpiece model with the point cloud data, calculating expected polishing amount of the robot polishing process according to the point cloud data, and generating a processing path, an expected pose and expected polishing force in the robot polishing process according to an inverse kinematics model of the robot polishing process.
And step two, simplifying the robot into a mass-damping-spring system under the condition of no impedance control, acquiring a plurality of position points of the robot in the Z direction of a workpiece coordinate system in the process of moving along the processing path, and constructing a target function according to the plurality of position points and key parameters of impedance control in the grinding and polishing processing of the robot.
Specifically, under the condition of no-resistance control, the robot is simplified into a mass-damping-spring system, the robot is controlled to move along a processing path, a plurality of position points of the robot in the z direction of a tool coordinate system in the work process of the robot are collected, polynomial fitting is carried out on the plurality of acquired position points, key parameters in an objective function are determined, and therefore the objective function of the primary constraint condition is determined. Wherein, the key parameters in the objective function are a mass matrix M in impedance control, a damping matrix D in impedance control and an elastic matrix K in impedance controlp. In the invention, least square method is adopted to carry out polynomial fitting. The calculation model of the objective function is as follows:
wherein N is the total number of position points continuously collected along the processing path, k is a position point sampling sequence index, fd(k) F (k) is a feedback value of the normal force of the robot at the single position point, and the calculation formula of the feedback value is as follows:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, KpIs an elastic matrix in impedance control, s (k),
Respectively displacement, velocity and acceleration of the robot in the normal direction of the workpiece.
That is, in the present invention, the square of the difference between the contact force and the expected force at all points on the whole machining path is selected as the objective function, and the objective function can be more representative of the whole effect on the machining path. Meanwhile, the invention carries out primary prepositive simulation on the actual machining contact force in the non-resistance environment and is used for guiding the feedback regulation of the actual grinding and polishing machining process in the later period, so that the feedback regulation in the actual machining is quicker and the precision is higher.
And step three, constructing a constraint condition of the key parameter, and acquiring the optimal solution of the key parameter by adopting a weight-improved particle swarm algorithm according to the constraint condition and the target function. M, D, K for robot system by using weight improved particle swarm algorithmpAnd optimizing the parameters to obtain an optimal solution of the impedance force control parameters for workpiece grinding and polishing, thereby optimizing a force control algorithm of the robot.
In the invention, constraint conditions are set according to adaptive conditions of control and adjustment of grinding and polishing processing impedance of a robot, specifically, a calculation model of the constraint conditions is as follows:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, KpIs an elastic matrix in impedance control.
In the weight-improved particle swarm optimization, an objective function is taken as a target of particle optimization, and a constraint condition is used for limiting the boundary of the particle optimization. Specifically, the method comprises the following steps:
(1) and acquiring the original data of the key parameters at each position point, and performing initialization processing to obtain randomly generated initialization particles and the positions and the speeds of the particles.
(2) Evaluating the fitness of each particle according to the objective function, and storing the position and the fitness value of the particle in the individual extreme value p of the particlebestIn (1), all individuals are extremized by pbestThe position of the particle with the optimal adaptive value and the adaptive value are stored in the global extreme value gbestIn (1).
(3) And updating the position and the speed of the particle according to the constraint condition. In the invention, the position and the speed of the particles are updated by adopting the following formula:
wherein, c1And c2Is the number of accelerations, r1、r2Two are in [0,1 ]]Acceleration weight coefficient, x, varying in betweeni,j(t) is the displacement value of the ith particle at time t, vi,j(t) is the velocity value of the ith particle at time t, w is the nonlinear dynamic inertia weight value, pi,jFor individual extrema of different particles, pg,jAnd d is the dimension of the search space.
Furthermore, the invention introduces a Particle Swarm in the weight-improved Particle Swarm algorithm (Particle Swarm Optimization,PSO) algorithm, setting the population size q of the particle as basic parameter and the threshold gen of evolution algebramaxCurrent evolution algebra gen, freely distributed over [0,1 ]]Random number r of interval1、r2. Setting a supplemental parameter acceleration constant c1And c2Maximum flying velocity v of particlemaxWherein c is1Maximum step size representing the best global particle for the flight, c2Maximum step size, v, representing the best particle to fly to an individualmaxReflecting the speed threshold value of each dimension of the particle flight, and adjusting the size of the particle to enable the searching capability of the particle population to be freely switched in a global/local mode. Generating an initial population in d-dimensional space, and assigning a particle population of size q as X ═ X1,x2,…,xi,…,xD) And establishing a three-dimensional vector scale of each particle individual, namely, any x is described by (x, V, P), wherein x represents a position, V represents a speed, and P represents an individual extreme value. For the ith particle, the position is xi,j(t) velocity vi,j(t) individual extremum is pi,j. Setting global extreme value of population as pg,jAnd carrying out three-dimensional assignment on each particle, carrying out speed and position iterative updating according to the following rule, calculating the individual extreme value and the global extreme value of each generation of particles, and preferably carrying out calculation iteration.
(4) And updating the inertia weight in the particle swarm algorithm. The inertia weight is updated according to the particle objective function value, when the particle objective function value is dispersed, the inertia weight is reduced, and when the particle objective function value is consistent, the inertia weight is increased. That is, for the particle updated in the iteration of the gen time, the objective function is calculated, the superiority and inferiority of the objective function are compared with those of the previous generation, the population algebra parameters are corrected after the speed and position of the particle in the current generation are updated, and when the iteration algebra reaches a specified threshold, the program is terminated. At this time, the final iterated calculation value is output, i.e. the optimal solution.
Updating the inertia weight by adopting a nonlinear dynamic inertia weight formula as follows:
wherein f represents a real-time objective function value of the particle; f. ofavgAnd fminRespectively representing the average value and the minimum target value of all the current particles, w is a nonlinear dynamic inertia weight value, wminIs the minimum value of the inertial weight, wmaxIs the maximum value of the inertial weight.
(5) Comparing the adaptive value of each particle with the best position of the particle, if the adaptive value does not meet the preset requirement, returning to the step S33, if the adaptive value meets the preset requirement, taking the current value as the best position of the particle, and comparing the individual extreme values p of all the current particlesbestAnd global extreme gbestAnd updating the current global extremum gbest。
(6) When the particle swarm algorithm reaches a stopping condition, stopping searching and outputting the optimal solution of the key parameters; otherwise, the step S33 is returned to continue the search.
And step four, updating and adjusting the pose and the grinding and polishing force of the grinding and polishing robot in real time according to the optimal solution of the key parameters, the expected grinding and polishing amount and the expected pose of the grinding and polishing robot, the actual pose of the robot and the actual grinding and polishing force in the grinding and polishing process of the robot, so that the force tracking error of the grinding and polishing process of the robot along the processing path is minimum, and further the grinding and polishing process of the workpiece to be processed with the expected grinding and polishing force is realized. The robot system monitors the change of grinding and polishing force in real time through the six-dimensional force sensor, and calls a force control algorithm according to the feedback of the machining force to adjust the pose of the robot, so that grinding and polishing with expected force are realized.
In the grinding and polishing process of the robot, the optimal solution of the key parameters is continuously updated to realize the force control in the grinding and polishing process of the robot, so that the pose and the grinding and polishing force of the grinding and polishing process of the robot are updated and adjusted in real time, and the calculation model of the motion differential equation of the robot in the force control process is as follows:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, and KpFor the elastic matrix in impedance control, x (t),
Respectively displacement, speed and acceleration of the robot in the normal direction of the workpiece, i is the initial distance between the workpiece and the grinding and polishing wheel, v is the initial speed when the workpiece moves towards the grinding and polishing wheel in the opposite direction, and KeIs an elastic matrix in contact with the environment.
According to the method, the square of the difference between the contact force of all position points on the whole machining path and the expected force is taken as a target function, the target function can represent the whole effect on the machining path better, secondly, the method performs one-time pre-simulation on the actual machining contact force under the impedance-free environment, the feedback lag time is shorter, the reaction is faster, the control is more accurate, and further, the tracking effect of the impedance control on the expected force is effectively improved; on the other hand, the processing precision of the grinding and polishing of the robot is effectively improved, and the surface consistency and good surface roughness of the processed workpiece are ensured.
Example 1
The active control method of the robot grinding and polishing machining contact force is used for improving the tracking effect of impedance control on expected force and improving the machining effect of a grinding and polishing system and the surface quality of a workpiece, and specifically comprises the following steps:
(1) offline programming and simulation are carried out on a blade model in RobotStudio software, and calibration of sensors such as a robot, a workpiece coordinate system, a scanner, a six-dimensional force sensor and the like is completed on site;
(2) acquiring coordinate values of the robot in the z direction of a tool coordinate system by utilizing the robot motion under the impedance-free control, performing polynomial fitting on the acquired position points, and determining key parameters in an objective function so as to determine the objective function of a primary constraint condition:
where N is the total number of samples in the continuous motion path, k is the sample order index, and fd(k) Expected values for the normal force of the robot at a single path point, f (k) fd(k) Feedback values for the normal force of the robot at a single path point.
(3) M, D, K for robotic system using weight-refined particle swarm optimizationpAnd optimizing the parameters to obtain an optimal solution of impedance control parameters for workpiece grinding and polishing, thereby optimizing a force control algorithm of the robot. The system parameters of the impedance control strategy are optimized by using a weight-improved particle swarm algorithm.
In the impedance control strategy, the robot is simplified into a mass-damping-spring system, the position or force is not directly controlled in the control process, but indirect force control is realized by adjusting mechanical impedance (including position impedance, speed impedance and acceleration impedance), and a motion differential equation is expressed as follows:
wherein M is a stiffness matrix of the system, D is a damping matrix of the system, KpIs a system elastic matrix, KeIs the elastic matrix of the system in contact with the environment.
The particle swarm algorithm with improved weight comprises the following basic steps:
1) randomly initializing the position and speed of each particle in the population;
2) evaluating the fitness of each particle, and storing the position and the fitness value of each particle into an individual extreme value p of each particlebestIn (1), all p arebestThe individual position of the optimum adaptive value and the adaptive value are stored in the global extreme value gbestPerforming the following steps;
3) updating the position and the speed of the particles;
the particle swarm algorithm with improved weight updates the particle position and velocity using the following formula:
xi,j(t+1)＝xi,j(t)+vi,j(t+1),j＝1,2,...,d
vi,j(t+1)＝w·vi,j(t)+c1r1[pi,j-xi,j(t)]+c2r2[pg,j-xi,j(t)]
in the formula, c1，c2Respectively representing the acceleration numbers, which are numbers of about 2 random frequently; r is1，r2Two are in [0,1 ]]The acceleration weight coefficients that vary between are randomly generated.
4) Updating the weight;
the particle swarm optimization with improved weight utilizes the following nonlinear dynamic inertia weight formula to update the weight:
wherein f represents a real-time objective function value of the particle; f. ofavgAnd fminRespectively, the average and minimum target values of all the particles at present.
In the particle swarm optimization nonlinear dynamic inertia weight formula, the inertia weight changes along with the change of the particle objective function value. Decreasing the inertial weight when the particle target value is dispersed; when the particle target values are consistent, the inertia weight is increased.
5) Of each particleThe adapted value is compared with the best position of the particle, and if so, the current value is taken as the best position of the particle. Comparing all current pbestAnd gbestAnd update gbest；
6) When the algorithm reaches the stop condition, stopping searching and outputting the result; otherwise, returning to 3) and continuing searching.
(4) In the field processing process, the robot clamps a workpiece, a scanner is used for scanning to obtain complete point cloud data of the workpiece, point cloud matching is carried out to obtain the correction amount of the workpiece on an original CAD model, and the result is output to the robot;
(5) the robot system monitors the change of grinding and polishing force in real time through the force sensor, and calls a force control algorithm according to the feedback of the processing force to adjust the pose of the robot, so that the grinding and polishing with expected force are realized.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (9)
1. An active control method for a contact force of grinding and polishing of a robot is characterized by comprising the following steps:
s1, calculating expected polishing amount and processing path of the robot polishing processing according to the appearance of the workpiece to be processed and the three-dimensional model of the workpiece, and generating expected pose and expected polishing force in the robot polishing processing process according to the inverse kinematics model of the robot polishing processing;
s2, simplifying the robot into a mass-damping-spring system under the condition of no impedance control, acquiring a plurality of position points of the robot in the Z direction of a workpiece coordinate system in the process of moving along the processing path, and constructing a target function according to the plurality of position points and key parameters of impedance control in the grinding and polishing processing of the robot;
s3, constructing a constraint condition of the key parameter, and acquiring an optimal solution of the key parameter by adopting a weight-improved particle swarm algorithm according to the constraint condition and an objective function;
and S4, updating and adjusting the pose and the grinding and polishing force of the grinding and polishing robot in real time according to the optimal solution of the key parameters, the expected grinding and polishing amount and the expected pose of the grinding and polishing robot, the actual pose of the robot and the actual grinding and polishing force in the grinding and polishing process of the robot, so that the force tracking error of the grinding and polishing process of the robot along the processing path is minimum, and further the grinding and polishing process of the workpiece to be processed with the expected grinding and polishing force is realized.
2. The active control method of the contact force of the robot grinding and polishing process of claim 1, wherein in step S1, the specific steps of calculating the expected grinding and polishing amount and the processing path of the robot grinding and polishing process according to the shape of the workpiece to be processed and the three-dimensional model of the workpiece are as follows:
s11, calibrating the robot, the workpiece coordinate system and the force sensor for measuring the actual grinding and polishing force;
s12, performing off-line programming and trajectory planning on the workpiece through the three-dimensional model of the workpiece to generate a processing path in the grinding and polishing process of the robot;
s13, point cloud data of the appearance of the workpiece to be processed is obtained through measurement, and the point cloud data are matched with the three-dimensional model of the workpiece, so that the expected polishing amount of the robot for polishing is obtained.
3. The active control method of contact force of robot grinding and polishing process of claim 1, wherein in step S2, the calculation model of the objective function is:
wherein N is the total number of position points continuously collected along the processing path, k is a position point sampling sequence index, fd(k) A desired value of the normal force of the robot at a single position point, f (k) a feedback value of the normal force of the robot at a single position point, a calculation formula of the feedback valueComprises the following steps:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, KpIs an elastic matrix in impedance control, s (k),
Respectively displacement, velocity and acceleration of the robot in the normal direction of the workpiece.
4. The active control method of contact force of robot grinding and polishing process of claim 1, wherein in step S3, the calculation model of the constraint function is:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, KpIs an elastic matrix in impedance control.
5. The active control method for the contact force of the robot grinding and polishing process of claim 1, wherein in step S4, the robot continuously updates the optimal solution of the key parameter to realize the force control during the grinding and polishing process of the robot, so as to update and adjust the pose and the grinding and polishing force of the robot grinding and polishing process in real time, and the calculation model of the robot motion differential equation during the force control process is:
wherein M is a mass matrix in impedance control, D is a damping matrix in impedance control, and KpFor the elastic matrix in impedance control, x (t),
Respectively displacement, speed and acceleration of the robot in the normal direction of the workpiece, i is the initial distance between the workpiece and the grinding and polishing wheel, v is the initial speed when the workpiece moves towards the grinding and polishing wheel in the opposite direction, and KeIs an elastic matrix in contact with the environment.
6. The active control method of the contact force of the robot grinding and polishing process according to claim 1, wherein the step S3 specifically comprises the following steps:
s31, collecting the original data of the key parameters at each position point, and carrying out initialization processing to obtain randomly generated initialization particles and the positions and speeds of the particles;
s32, evaluating the fitness of each particle according to the objective function, and storing the position and the fitness value of the particle in the individual extreme value p of the particlebestIn (1), all individuals are extremized by pbestThe position of the particle with the optimal adaptive value and the adaptive value are stored in the global extreme value gbestPerforming the following steps;
s33, updating the position and the speed of the particle according to the constraint condition;
s34, updating the inertia weight in the particle swarm algorithm;
s35, comparing the adaptive value of each particle with the best position of the particle, if the adaptive value does not meet the preset requirement, returning to the step S33, if the adaptive value meets the preset requirement, taking the current value as the best position of the particle, and comparing the individual extreme values p of all the current particlesbestAnd global extreme gbestAnd update whenFormer global extreme gbest；
S36, when the particle swarm algorithm reaches a stop condition, stopping searching and outputting the optimal solution of the key parameters; otherwise, the step S33 is returned to continue the search.
7. The active control method of contact force in robot polishing and burnishing process of claim 6, wherein in step S33, the position and velocity of the particles are updated according to the following formula:
wherein, c1And c2Is the number of accelerations, r1、r2Two are in [0,1 ]]Acceleration weight coefficient, x, varying in betweeni,j(t) is the displacement value of the ith particle at time t, vi,j(t) is the velocity value of the ith particle at time t, w is the nonlinear dynamic inertia weight value, pi,jFor individual extrema of different particles, pg,jAnd d is the dimension of the search space.
8. The active control method of contact force in grinding and polishing process of robot as claimed in claim 6, wherein in step S34, the inertia weight is updated by using the following nonlinear dynamic inertia weight formula:
wherein f represents a real-time objective function value of the particle; f. ofavgAnd fminRespectively representing the average value and the minimum target value of all the current particles, w is a nonlinear dynamic inertia weight value, wminIs the minimum value of the inertial weight, wmaxIs the maximum value of the inertial weight.
9. The active control method of contact force in robot polishing and grinding process of claim 6, wherein in step S34, the inertia weight is updated according to the objective function value of the particle, and when the objective function value of the particle is dispersed, the inertia weight is decreased, and when the objective function value of the particle is consistent, the inertia weight is increased.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011547920.7A CN112666831A (en) | 2020-12-24 | 2020-12-24 | Active control method for grinding and polishing contact force of robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011547920.7A CN112666831A (en) | 2020-12-24 | 2020-12-24 | Active control method for grinding and polishing contact force of robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112666831A true CN112666831A (en) | 2021-04-16 |
Family
ID=75409891
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011547920.7A Pending CN112666831A (en) | 2020-12-24 | 2020-12-24 | Active control method for grinding and polishing contact force of robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112666831A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113319863A (en) * | 2021-05-11 | 2021-08-31 | 华中科技大学 | Workpiece clamping pose optimization method and system for grinding and polishing blisk by robot |
| CN113459093A (en) * | 2021-06-22 | 2021-10-01 | 天津理工大学 | Impedance control method of polishing robot |
| CN113561168A (en) * | 2021-06-11 | 2021-10-29 | 华中科技大学 | Speed reduction and torque increase control device, method and equipment based on torque control and storage medium |
| CN113894798A (en) * | 2021-11-26 | 2022-01-07 | 遨博（北京）智能科技有限公司 | Robot control method and robot system |
| CN113910085A (en) * | 2021-09-26 | 2022-01-11 | 江苏集萃华科智能装备科技有限公司 | Control method, control device and control system of metal mirror polishing production line |
| CN114290335A (en) * | 2022-01-17 | 2022-04-08 | 河南科技大学 | Robot track planning method |
| CN114346851A (en) * | 2021-12-06 | 2022-04-15 | 襄阳华中科技大学先进制造工程研究院 | Blade grinding process parameter adjusting method, device, equipment and storage medium |
| CN114434449A (en) * | 2022-04-02 | 2022-05-06 | 北京科技大学 | Novel particle swarm adaptive impedance control method and device |
| CN114800505A (en) * | 2022-04-25 | 2022-07-29 | 珠海格力智能装备有限公司 | Robot operation control method and device |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107274012A (en) * | 2017-06-06 | 2017-10-20 | 上海电机学院 | Short-term wind power forecast method based on cloud evolution particle cluster algorithm |
| CN107729706A (en) * | 2017-11-29 | 2018-02-23 | 湖南科技大学 | A kind of kinetic model construction method of Nonlinear Mechanical Systems |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN109434843A (en) * | 2018-12-10 | 2019-03-08 | 华中科技大学 | A kind of device and method of the Robot Force console keyboard mill blade based on dragging teaching |
| CN109816119A (en) * | 2019-02-02 | 2019-05-28 | 南京邮电大学 | A kind of wireless sensor abnormal deviation data examination method based on unsupervised learning |
| CN109884900A (en) * | 2019-04-03 | 2019-06-14 | 东南大学 | The design method of cropper path following control device based on adaptive model PREDICTIVE CONTROL |
| CN111752151A (en) * | 2020-06-17 | 2020-10-09 | 华中科技大学 | Adaptive tracking and compensating method and system for grinding and polishing industrial blade |
| CN111823100A (en) * | 2020-07-24 | 2020-10-27 | 前元运立(北京)机器人智能科技有限公司 | Robot-based small-curvature polishing and grinding method |
-
2020
- 2020-12-24 CN CN202011547920.7A patent/CN112666831A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107274012A (en) * | 2017-06-06 | 2017-10-20 | 上海电机学院 | Short-term wind power forecast method based on cloud evolution particle cluster algorithm |
| CN107729706A (en) * | 2017-11-29 | 2018-02-23 | 湖南科技大学 | A kind of kinetic model construction method of Nonlinear Mechanical Systems |
| CN108120451A (en) * | 2017-12-21 | 2018-06-05 | 苏州大学 | Based on silicon micro accerometer temperature-compensation method, the system for improving PSO optimization neural networks |
| CN109434843A (en) * | 2018-12-10 | 2019-03-08 | 华中科技大学 | A kind of device and method of the Robot Force console keyboard mill blade based on dragging teaching |
| CN109816119A (en) * | 2019-02-02 | 2019-05-28 | 南京邮电大学 | A kind of wireless sensor abnormal deviation data examination method based on unsupervised learning |
| CN109884900A (en) * | 2019-04-03 | 2019-06-14 | 东南大学 | The design method of cropper path following control device based on adaptive model PREDICTIVE CONTROL |
| CN111752151A (en) * | 2020-06-17 | 2020-10-09 | 华中科技大学 | Adaptive tracking and compensating method and system for grinding and polishing industrial blade |
| CN111823100A (en) * | 2020-07-24 | 2020-10-27 | 前元运立(北京)机器人智能科技有限公司 | Robot-based small-curvature polishing and grinding method |
Non-Patent Citations (2)
| Title |
|---|
| 丰飞等: "大型风电叶片多机器人协同磨抛系统的设计与研究", 《机器人技术与应用》 * |
| 周亚军: "基于多维力传感器的机器人叶片磨抛力控制技术研究", 《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113319863B (en) * | 2021-05-11 | 2023-06-16 | 华中科技大学 | Workpiece clamping pose optimization method and system for robot grinding and polishing machining of blisk |
| CN113319863A (en) * | 2021-05-11 | 2021-08-31 | 华中科技大学 | Workpiece clamping pose optimization method and system for grinding and polishing blisk by robot |
| CN113561168A (en) * | 2021-06-11 | 2021-10-29 | 华中科技大学 | Speed reduction and torque increase control device, method and equipment based on torque control and storage medium |
| CN113459093A (en) * | 2021-06-22 | 2021-10-01 | 天津理工大学 | Impedance control method of polishing robot |
| CN113910085B (en) * | 2021-09-26 | 2022-11-18 | 江苏集萃华科智能装备科技有限公司 | Control method, control device and control system of metal mirror polishing production line |
| CN113910085A (en) * | 2021-09-26 | 2022-01-11 | 江苏集萃华科智能装备科技有限公司 | Control method, control device and control system of metal mirror polishing production line |
| CN113894798A (en) * | 2021-11-26 | 2022-01-07 | 遨博（北京）智能科技有限公司 | Robot control method and robot system |
| CN114346851A (en) * | 2021-12-06 | 2022-04-15 | 襄阳华中科技大学先进制造工程研究院 | Blade grinding process parameter adjusting method, device, equipment and storage medium |
| CN114290335A (en) * | 2022-01-17 | 2022-04-08 | 河南科技大学 | Robot track planning method |
| CN114290335B (en) * | 2022-01-17 | 2023-08-08 | 河南科技大学 | Robot track planning method |
| CN114434449B (en) * | 2022-04-02 | 2022-08-09 | 北京科技大学 | Novel particle swarm adaptive impedance control method and device |
| CN114434449A (en) * | 2022-04-02 | 2022-05-06 | 北京科技大学 | Novel particle swarm adaptive impedance control method and device |
| CN114800505A (en) * | 2022-04-25 | 2022-07-29 | 珠海格力智能装备有限公司 | Robot operation control method and device |
| CN114800505B (en) * | 2022-04-25 | 2024-02-20 | 珠海格力智能装备有限公司 | Robot operation control method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112666831A (en) | Active control method for grinding and polishing contact force of robot | |
| CN111964575B (en) | Digital twin modeling method for milling of mobile robot | |
| CN109782601B (en) | Design method of self-adaptive neural network synchronous robust controller of coordinated mechanical arm | |
| JP7199178B2 (en) | ROBOT CONTROL DEVICE, ROBOT DEVICE, ROBOT CONTROL PARAMETER ADJUSTMENT METHOD, AND PROGRAM | |
| CN110948504B (en) | Normal constant force tracking method and device for robot machining operation | |
| CN105817712A (en) | Scraping device and scraping method using robot | |
| CN110315396B (en) | Industrial robot constant-force grinding and polishing method based on big data | |
| CN111975579B (en) | Robot constant-force polishing system based on polishing model and iterative algorithm | |
| CN114474039B (en) | Method for supporting adjustment of parameter set of robot and information processing apparatus | |
| CN111752151A (en) | Adaptive tracking and compensating method and system for grinding and polishing industrial blade | |
| CN111596614A (en) | Motion control error compensation system and method based on cloud edge cooperation | |
| Latifinavid et al. | High-performance parallel hexapod-robotic light abrasive grinding using real-time tool deflection compensation and constant resultant force control | |
| CN114310914A (en) | Fuzzy self-adaptive iterative trajectory tracking control method and system for multi-degree-of-freedom mechanical arm | |
| Pham et al. | Hierarchical PID controller design for a flexible link robot manipulator using the bees algorithm | |
| CN114660940A (en) | Grinding constant force control method based on deep reinforcement learning PPO algorithm | |
| CN112327938B (en) | Robot near-zero following error control method based on data driving | |
| Dai et al. | Adaptive friction compensation control of robotic pneumatic end-effector based on LuGre model | |
| CN112847362B (en) | Transfer robot vision servo control method based on image moment | |
| CN114859698A (en) | Trajectory tracking control method and device of 6-3-PUS parallel mechanism | |
| CN111203883B (en) | Self-learning model prediction control method for robot electronic component assembly | |
| CN110788859B (en) | Controller parameter universe self-adaptive adjustment system | |
| CN114248267A (en) | Parameter display method, storage medium, and information processing apparatus | |
| CN113848808A (en) | Tool path fairing and optimization method based on measurement point cloud | |
| CN113618728A (en) | Method and device for compensating motion trail of robot and computer storage medium | |
| CN114367974B (en) | Method for adjusting parameter set of robot, recording medium, and information processing device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210416 |
|
| RJ01 | Rejection of invention patent application after publication |