WO2020237407A1

WO2020237407A1 - Method and system for self-calibrating robot kinematic parameter, and storage device

Info

Publication number: WO2020237407A1
Application number: PCT/CN2019/088251
Authority: WO
Inventors: 李康宁
Original assignee: 深圳配天智能技术研究院有限公司
Priority date: 2019-05-24
Filing date: 2019-05-24
Publication date: 2020-12-03
Also published as: CN112105484A; CN112105484B

Abstract

A method for self-calibrating a robot kinematic parameter, comprising: a calibrated robot enables an end effector to contact with the surface of a standard workpiece for multiple times, and records angles of axes of the calibrated robot during each contact, the standard workpiece being mounted on a flange of a reference robot; the reference robot changes the attitude of the standard workpiece; the calibrated robot repeatedly executes the step of enabling the end effector to contact with the surface of the standard workpiece for multiple times, and recording the angles of axes of the calibrated robot during each contact so as to obtain a plurality of groups of axis angle values of the calibrated robot; and obtain the actual kinematic parameter of the calibrated robot according to the plurality of groups of axis angle values, the nominal kinematic parameter of the calibrated robot, and the actual diameter of the standard workpiece. The method can improve the efficiency, achieve automatic calibration, and make it easy to achieve batch calibration. The present invention also relates to a system for executing the method, and a storage device for storing a program file of the method.

Description

Robot kinematics parameter self-calibration method, system and storage device

【Technical Field】

This application relates to the field of robotics, and in particular to a method, system and storage device for self-calibration of robot kinematics parameters.

【Background technique】

The motion accuracy of the robot plays a vital role in the application reliability of industrial robots in production. The geometric parameter (DH parameter) error of each connecting rod of the robot is the main link that causes the robot system error, which is mainly caused by the deviation between the actual geometric parameters of the connecting rod and the theoretical parameter value during the manufacturing and installation process.

There are two commonly used robot DH parameter calibration methods: one is to use external measurement equipment such as laser tracker or three-coordinate measuring machine to measure the pose of the calibrated robot TCP point at each measurement point, according to the actual measurement of the pose and the use of the robot The error of the theoretical measurement pose obtained by the positive solution of the nominal DH parameter is used to identify the error of the DH parameter. However, due to the characteristics of high cost and large volume of large-scale high-precision measurement equipment such as laser trackers, it is difficult to meet the demand for simultaneous calibration of multiple robots to improve efficiency during mass production of robots.

Another method is to use the self-calibration method of standard workpieces. The machine tool probe is installed on the robot flange, and the special calibration standard workpieces are three standard balls fixed on the base. During the calibration process, the robot is controlled so that the machine tool probe touches The surface of each standard ball is changed many times, the pose of the standard ball workpiece base is repeated, the above measurement process is repeated, and the Cartesian position coordinate of the measurement point on the surface of the standard ball under the nominal DH parameter is calculated according to the measured shaft position and the positive kinematics solution, and Fit multiple sets of sphere centers through spherical fitting, and calculate the distance of theoretical sphere centers. The DH parameters are calibrated using the calculated error between the theoretical sphere center distance and the actual sphere center distance. However, a custom-made workpiece with three standard balls is used as a fixed physical constraint in self-calibration, and a standard workpiece with high machining accuracy is required, and an additional solution for adjusting the pose of the workpiece is required, which increases the cost and reduces the degree of automation.

[Content of the invention]

This application proposes a method, system and storage device for self-calibration of robot kinematics parameters, which can reduce costs, can realize automatic calibration, and is easy to realize batch calibration.

In order to solve the above problems, this application proposes a robot kinematic parameter self-calibration method, which includes: the calibration robot contacts the end effector multiple times to the surface of the standard workpiece, and records the angle of each axis of the calibration robot each time. The standard workpiece is installed on the flange of a reference robot; the reference robot changes the pose of the standard workpiece, the calibration robot repeatedly executes the end effector contacting the surface of the standard workpiece multiple times, and records each contact The step of calibrating the angles of each axis of the robot to obtain multiple sets of axis angle values of the calibration robot; using the multiple sets of axis angle values, the nominal kinematic parameters of the calibration robot, and the actual radius of the standard workpiece to obtain The actual kinematics parameters of the calibration robot.

In order to solve the above problems, this application also proposes a robot kinematic parameter self-calibration system, including: a calibration robot and a reference robot; the flange of the calibration robot is equipped with an end effector, and the flange of the reference robot is equipped with Standard workpiece; the reference robot is used to change the pose of the standard workpiece; the calibration robot is used to execute instructions to implement the above-mentioned robot kinematic parameter self-calibration method.

In order to solve the above problems, this application also proposes a storage device with a program file stored inside, which is characterized in that the program file is executed to realize the above-mentioned robot kinematic parameter self-calibration method.

Different from the prior art, in the embodiment of the present application, the reference robot with standard workpieces mounted on the flange is used to assist the calibration. The calibration tool is simple and efficient. During the calibration, only the movement of the calibration robot and the reference robot needs to be input once. In the process of calibration, the calibration robot and the reference robot can automatically move and calculate without manual intervention, so the calibration has a high degree of automation, which is suitable for large-scale calibration requirements in the robot production process.

【Explanation of drawings】

FIG. 1 is a schematic flowchart of a first embodiment of a method for self-calibration of robot kinematics parameters according to the present application;

Fig. 2 is a schematic diagram of an execution scenario of the method shown in Fig. 1;

FIG. 3 is a schematic diagram of a specific flow of steps S11 and S12 in FIG. 1;

FIG. 4 is a schematic diagram of a specific flow of step S13 in FIG. 1;

FIG. 5 is a schematic flowchart of a second embodiment of a method for self-calibration of robot kinematic parameters according to the present application;

FIG. 6 is a schematic diagram of the specific flow of each step in FIG. 5;

Fig. 7 is a schematic structural diagram of a first embodiment of a robot kinematic parameter self-calibration system according to the present application;

8 is a schematic structural diagram of a second embodiment of the robot kinematic parameter self-calibration system according to the present application;

FIG. 9 is a schematic structural diagram of an embodiment of a storage device of the present application.

【Detailed ways】

The application will be described in detail below with reference to the drawings and embodiments.

As shown in Fig. 1, the first embodiment of the robot kinematic parameter self-calibration method of the present application includes:

S11: The calibration robot touches the end effector to the surface of the standard workpiece multiple times, and records the angle of each axis of the calibration robot during each contact.

Among them, the end effector is installed on the flange of the calibration robot, and the standard workpiece is installed on the flange of the reference robot. The angle of each axis of the robot refers to the angle of rotation of each joint of the robot with each joint axis as the rotation axis.

As shown in Figure 1 and Figure 2, in an application example, the flange of the calibration robot A is equipped with an end effector 10, and the flange of the reference robot B is equipped with a standard workpiece 20, so that the calibration robot A can touch the standard workpiece At different positions on the surface of 20, the calibration robot A and the reference robot B are arranged relative to each other, so that the smart working spaces of the calibration robot A and the reference robot B have a larger overlap, which means that the moving spaces of the two manipulators overlap more. , For example, the coincidence degree reaches at least 50%. The end effector 10 can be the side head of the machine tool as shown in Fig. 2. In order to meet the requirements of calibration accuracy, the machine tool probe can choose a model with a repeatability of 1um and below, and the standard workpiece 20 can be a standard ball workpiece. Need to have a higher degree of sphericity and surface smoothness, such as the three-coordinate gauge standard ball.

During calibration, the reference B controlled robot manipulator, so that the position and orientation of the workpiece 20 is a standard C _i, which control the calibration of the robot manipulator A, the end effector 10 may be such that multiple contact standard surface 20 of the workpiece, and are in contact with each Record the angle of each axis of the calibration robot A. Among them, the calibration robot A's own controller (not shown) records the angles of each axis of the calibration robot A after the end effector 10 touches the surface of the standard workpiece 20. Of course, the calibration robot A can also control the axis The angle is transmitted to an external device or related files are generated for storage and recording.

Optionally, as shown in FIG. 3, step S11 specifically includes:

S111: The end effector of the calibration robot touches different positions on the surface of the standard workpiece m times in different postures, and records the angle of each axis of the calibration robot each time it touches.

In this embodiment, the calibration robot is a chain robot. The chain robot can be described by the structure of joint-link-joint-link -...-link-end effector. DH parameters (Denavit–Hartenberg parameters) are the four parameters of the mechanical arm mathematical model and coordinate system that are proposed by Denavit and Hartenberg in 1955 to express the position and angle relationship of adjacent joints and the links between them.

Since the number of DH parameters corresponding to each link of the chain robot is 4, in order to calibrate the kinematic parameters DH parameters of the robot more accurately, the number of times m that the calibration robot A contacts the standard workpiece 20 in the same pose should be greater than 4, for example Default m=8.

Specifically, in conjunction with Figures 2 and 3, in order to improve the accuracy of calibration, when the calibration robot A controls the end effector 10 to contact the standard workpiece 20, it can use different postures to contact different positions on the surface of the standard workpiece 20, and each time During contact, record the angle q _{ij of} each axis of the calibration robot A. For example, each time the end effector 10 touches the surface of the standard workpiece 20, the position is different, and the contact is made in a different pose, or the same position on the surface of the standard workpiece 20 is contacted with different poses some times.

S12: Refer to the robot to change the pose of the standard workpiece, and the calibration robot repeats step S11 to obtain multiple sets of axis angle values of the calibration robot.

Specifically, in order to obtain more sufficient data, the reference robot can change the pose of the standard workpiece multiple times. After each pose change, the calibration robot can repeat the above step S11, so that the calibration corresponding to the different poses of the standard workpiece can be obtained. The angle of each axis of the robot. The angle of each axis of the calibration robot recorded under the same pose of the standard workpiece forms a set of axis angle values, and then multiple sets of axis angle values of the calibration robot can be obtained for subsequent calculations of the calibration robot The actual kinematics parameters are actual DH parameters.

As shown in FIG. 2, the reference robot B can control the movement of its manipulator to change the pose of the standard workpiece 20, for example, from the pose C ₁ to the pose C ₂ , and then the calibration robot A can repeat step S11, namely The end effector 10 is controlled to contact different positions on the surface of the standard workpiece 20 for multiple times, and then two sets of axis angle values of the calibration robot corresponding to the pose C ₁ and the pose C _{2 of the} standard workpiece can be obtained.

Optionally, as shown in FIG. 3, step S12 includes:

S121: Determine whether the number of times the reference robot changes the pose of the standard workpiece is greater than (n-1).

Among them, in order to fully obtain the calibration data, the specific value of n is selected according to the number of axes of the calibration robot and the number of DH parameters of each axis. For example, when the calibration robot is a six-axis serial robot, the value of n should be greater than 20, such as n=60.

If the judgment result is not greater than, the following step S122 is executed, otherwise, the step S124 is executed.

S122: Change the pose of the standard workpiece, and return to step S111.

S124: Calibrate the angles of each axis of the robot according to the acquired n*m groups, and continue to perform step S13.

Specifically, in an application example, the reference robot uses a counting device, and every time it controls the manipulator to change the pose of the standard workpiece, the counting device counts once, then the reference robot changes the pose of the standard workpiece every time. , You can first determine whether the number of times the reference robot has changed the pose of the standard workpiece meets the requirements, that is, whether the number of changes has been greater than (n-1) (such as whether it is greater than 50), that is, whether the standard workpiece has been completed in n poses Measure multiple sets of angles of each axis of the calibration robot. If the number of changes is less than (n-1), you can continue to perform step S122 at this time to change the pose of the standard workpiece, and return to step S111 to obtain the calibration robot corresponding to the changed pose of the standard workpiece. The angle of each axis. If the number of changes has been greater than n, that is, the number of changes in the pose of the standard workpiece has reached the requirement, and the calibration data has been fully obtained at this time, and the subsequent step S13 can be continued.

S13: Calculate the actual kinematics parameters of the calibration robot by using the multiple sets of axis angle values, the nominal kinematics parameters of the calibration robot and the actual radius of the standard workpiece.

Among them, the standard workpiece is a standard spherical workpiece, the actual radius of which can be obtained in advance, and the nominal kinematics parameters of the calibration robot are usually preset at the factory or can be obtained in advance.

Specifically, using the forward and inverse solutions of the robot kinematics and the DH modeling method, the multiple sets of axis angle values, the nominal kinematic parameters of the calibration robot, the actual radius of the standard workpiece, and the actual kinematic parameters of the calibration robot can be established in advance. The relationship equation, so that the obtained multiple sets of axis angle values, the nominal kinematic parameters of the calibration robot, and the actual radius of the standard workpiece are substituted into the established relationship equation, and the relationship equation is solved, then the actual motion of the calibration robot can be calculated Learn parameters.

Optionally, as shown in FIG. 4, step S13 includes:

S131: Using the multiple sets of axis angle values and the nominal kinematic parameters of the calibration robot, the theoretical position of the contact point between the end effector and the surface of the standard workpiece corresponding to each set of axis angle values is calculated.

The positive solution of robot kinematics refers to the process of using the robot's nominal kinematic parameters (namely the nominal DH parameters) to solve the pose of the robot flange center in the Cartesian space robot base coordinate system with the known angle of each axis of the robot. When the end effector is installed on the flange, it is only necessary to perform a simple coordinate transformation on the calculated pose of the flange center in the robot base coordinate system to obtain the pose of the end effector in the robot base coordinate system. Therefore, by using the obtained multiple sets of axis angle values and nominal DH parameters of the calibration robot, according to the positive kinematics of the robot, the tool center point (ie TCP point) of the end effector of the calibration robot corresponding to each set of axis angle values can be calculated in the robot The pose in the base coordinate system. Since the nominal DH parameters are used in the calculation process, the calculated pose of the TCP point in the robot base coordinate system is the theoretical pose of the contact point between the end effector and the standard workpiece surface in the robot base coordinate system. The theoretical pose can get the theoretical position of the contact point of the end effector and the surface of the standard workpiece in the robot base coordinate system.

Specifically, when the acquired multiple sets of axis angle values of the calibration robot include n*m sets of calibration robot angles corresponding to n standard workpiece poses, the theory of n sets of contact points can be calculated according to the positive solution of the robot kinematics Position, each standard workpiece pose corresponds to the theoretical position of a group of contact points, the theoretical position of each group of contact points includes the theoretical position of m contact points, where the theoretical position of the contact point is the contact point in the robot base coordinate system Theoretical location.

S132: Using multiple sets of theoretical positions of contact points, fitting to obtain a fitting radius of the standard workpiece corresponding to each set of theoretical positions of contact points.

Specifically, when the standard workpiece is a standard spherical workpiece, using the calculated theoretical positions of multiple sets of contact points, the spherical fitting algorithm can be used to fit the fitting radius of the standard workpiece corresponding to the theoretical position of each set of contact points. The theoretical position of the group contact point can be fitted to get the fitting radius of a standard workpiece.

S133: Calculate actual kinematics parameters according to the fitted radius and the fitted radius error between the actual radius.

Among them, the actual radius of the standard workpiece can be obtained in advance after the standard workpiece to be used is selected, and the fitting radius error can be obtained by using the difference between the fitted radius and the actual radius.

Since the original data used in the fitting is the theoretical position of the contact point, and the theoretical position of the contact point is calculated using the nominal DH parameter, the fitting error between the fitted radius and the actual radius obtained by the fitting is the DH parameter error There is a correspondence between them, and by using this correspondence, the DH parameter error can be calculated, that is, the kinematic parameter error between the actual kinematic parameter (actual DH parameter) and the nominal kinematic parameter (nominal DH parameter). Then, since the nominal DH parameters of the calibration robot can be obtained in advance, the actual DH parameters of the calibration robot can be obtained by adding the kinematic parameter error to the nominal DH parameters.

In this embodiment, a reference robot with a standard workpiece mounted on the flange is used to assist the calibration. The calibration tool is simple and efficient. During calibration, you only need to input the motion instructions of the calibration robot and the reference robot once. During the calibration process, the robot and The reference robot can automatically move and calculate without manual intervention, so the calibration has a high degree of automation, which is suitable for large-scale calibration requirements in the robot production process.

As shown in FIG. 5, the second embodiment of the robot kinematic parameter self-calibration method of the present application is based on the first embodiment of the robot kinematic parameter self-calibration method of the present application, and before step S133 is further defined, it includes:

S201: Establish a first linear relationship equation between the position error of the contact point and the kinematic parameter error.

Among them, since the theoretical position of the contact point is calculated using the nominal kinematic parameters, the position error of the contact point is caused by the kinematic parameter error. The first difference between the contact point and the position error and the kinematic parameter error can be established. Linear relationship equation.

Since the linear relationship equation can be expressed by using a matrix equation, the solution of the linear relationship equation can be obtained conveniently and quickly by solving the matrix equation. Therefore, as shown in FIG. 6, step S201 may include:

S2011: Use the difference between the theoretical position and the actual position of the contact point as an element to establish an error matrix of the position of the contact point.

Among them, the theoretical position of the contact point is the theoretical position of the contact point in the robot base coordinate system, and the actual position of the contact point is the actual position of the contact point in the robot base coordinate system.

According to the robot's positive solution P _ij =f(q _ij ,DH _n ), where P _ij is the position of the end effector of the robot, q _ij is the angle of each axis of the robot, and DH _n is the nominal DH parameter. Using the recorded multiple sets of angle values of each axis of the calibration robot, the theoretical position p′ _ij of the contact point between the end effector of the calibration robot and the surface of the standard workpiece under the nominal DH parameter DH _n can be obtained. Assuming that the actual position of the contact point is p _ij , the difference between the theoretical position p′ _{ij of} the contact point and the actual position p _ij is Δp _ij =p′ _ij -p _ij , and then, the theoretical and actual positions of the contact point Difference as an element, the error matrix of the position of the contact point Δp=(Δp ₁₁ … Δp _nm ) ^T can be established, where n is the n poses of the standard workpiece used when recording the angle of each axis of the calibration robot, and each standard workpiece Each pose corresponds to m recorded positions of the angle of each axis of the calibration robot.

S2012: Use the difference between the nominal kinematics parameter and the actual kinematics parameter as the element to establish the error matrix of the kinematics parameter.

Assuming that the actual DH parameter is DH _a , and the nominal DH parameter of the calibration robot obtained is DH _n , since there are multiple DH parameters of the robot, both DH _a and DH _n can be expressed in a matrix, and the kinematic parameters of the calibration robot can be established The error matrix ΔDH=DH _n -DH _a .

S2013: Obtain the first Jacobian matrix of the theoretical position of the contact point with respect to the kinematic parameters.

Specifically, the Jacobian matrix is a matrix composed of first-order derivatives, and the theoretical position of the contact point with respect to the first Jacobian matrix J _DH of the kinematic parameters can be expressed as follows:

Among them, k is the number of identifiable DH parameters, and its specific value is related to the number of axes of the calibration robot. When a six-axis tandem robot is used in this embodiment, since the four related parameters of the first axis are not identifiable, k can be equal to 20. x ₁₁ ……z _nm are the coordinates of the theoretical position of the contact point obtained by the positive solution of the nominal DH parameter using the recorded axis angle value; DH ₁ ……DH _k is the nominal DH parameter.

S2014: Establish the first linear relationship equation shown in the following formula (1):

Δp=J _DH ΔDH (1)

Specifically, according to the kinematics principle of the robot, it can be deduced that there is an approximately linear relationship between the DH parameter error and the position error of the contact point as shown in the above formula (1). Among them, Δp represents the error matrix of the position of the contact point, J _DH represents the first Jacobian matrix, and ΔDH represents the error matrix of the kinematic parameters.

S202: Establish a second linear relationship equation between the fitting radius error of the standard workpiece and the position error of the contact point.

Among them, since the fitting radius of the standard workpiece is fitted by the theoretical position of the contact point, the fitting radius error of the standard workpiece is caused by the position error of the contact point, and the fitting radius error and contact of the standard workpiece can be established The second linear relationship equation between the position errors of the points.

Since the linear relation equation can be expressed by using a matrix equation, the solution of the linear relation equation can be solved conveniently and quickly by solving the matrix equation. Therefore, as shown in FIG. 6, step S202 may include:

S2021: Use the difference between the fitting radius of the standard workpiece and the actual radius as an element to establish a fitting radius error matrix of the standard workpiece.

Specifically, when the standard workpiece is a standard spherical workpiece, the theoretical positions of multiple sets of contact points are used for spherical fitting, that is, multiple fitting radii r _{i of} the standard workpiece can be obtained, and the actual radius r of the standard workpiece is selected The standard workpiece can be obtained in advance, and the difference between multiple sets of fitting radius r _i and the actual radius r Δr _i =r _i -r can be used as the element to establish the fitting radius error matrix of the standard workpiece Δr = (Δr ₁ … Δr _n ) ^T , n is the n poses of the standard workpiece used when recording the angle of each axis of the robot.

S2022: Obtain the second Jacobian matrix of the fitting radius of the standard workpiece with respect to the theoretical position of the contact point.

Specifically, the Jacobian matrix is a matrix composed of first-order derivatives, and the second Jacobian matrix J _p of the fitting radius of the standard workpiece with respect to the theoretical position of the contact point can be expressed as follows:

Among them, r ₁ ……r _n is the fitting radius obtained by fitting; x ₁₁ ……z _nm is the coordinate of the theoretical position of the contact point obtained by using the recorded n*m set of axis angle values through the nominal DH parameter positive solution.

S2023: Establish a second linear relationship equation shown in the following formula (2):

Δr=J _p Δp (2)

Specifically, it can be deduced that there is an approximately linear relationship between the fitting radius error and the position error of the contact point as shown in the above formula (2). Among them, Δr represents the fitting radius error matrix of the standard workpiece, J _p represents the second Jacobian matrix, and Δp represents the error matrix of the position of the contact point.

S203: Combine the first linear relationship equation and the second linear relationship equation to establish a third linear relationship equation between the fitting radius error and the kinematic parameter error.

Specifically, since the fitting radius error can be calculated, and there is a second linear relationship equation between the fitting radius error and the position error of the contact point, and there is a first linear relationship between the position error of the contact point and the kinematic parameter error Equation, therefore combining the first linear relationship equation and the second linear relationship equation, the third linear relationship equation between the fitting radius error and the kinematic parameter error can be established.

Optionally, when the first linear relationship equation is the above formula (1) and the second linear relationship equation is the above formula (2), as shown in FIG. 6, step S203 includes:

S2031: Substituting the first linear relationship equation into the second linear relationship equation to obtain the third linear relationship equation shown in the following formula (3):

Δr=JΔDH (3)

Among them, J=J _p J _DH .

The above steps S201 to S203 only need to be executed once, and the execution process only needs to be performed before step S133. After the third linear relationship equation is obtained, the required parameters can be calculated by substituting corresponding data.

Continue to refer to FIG. 5, step S133 specifically includes:

S204: Substitute the fitting radius error into the third linear relationship equation, and solve the third linear relationship equation to obtain actual kinematic parameters of the calibration robot.

After the third linear relationship equation and the fitting radius error are obtained, the fitting radius error can be substituted into the third linear relationship equation, and the third linear relationship equation can be solved, then the kinematic parameter error of the calibration robot can be calculated . Then, since the nominal kinematic parameters of the calibration robot can be obtained in advance, the actual kinematic parameters of the calibration robot can be obtained by superimposing the error of the kinematic parameter with the nominal kinematic parameters.

Among them, when there are many parameters, the linear equation can be solved conveniently and quickly by matrix operation. Therefore, as shown in FIG. 6, step S204 may include:

S2041: Solve the third linear relation equation to obtain the error matrix of the kinematic parameters of the calibration robot.

Among them, the error matrix of the kinematic parameters of the calibration robot is a matrix formed by the errors between the nominal DH parameters of the calibration robot and the actual DH parameters.

S2042: Superimpose the nominal kinematic parameters of the calibration robot and the error matrix of the kinematic parameters to obtain the actual kinematic parameters of the calibration robot.

Specifically, the third linear relationship equation can adopt the matrix equation shown in formula (3). In this case, the matrix inverse operation can be used to obtain the solution of the third linear relationship equation, that is, the movement of the calibration robot can be obtained. Then, according to the error matrix of the kinematic parameters and the nominal kinematic parameters of the calibration robot, the actual kinematic parameters of the calibration robot can be obtained.

In this embodiment, after the third linear relationship equation between the fitting radius error and the kinematic parameter error is established in advance, it is only necessary to substitute the fitting radius error into the third linear relationship equation after the fitting radius error is obtained. Using the matrix inverse operation, the actual kinematic parameters of the calibration robot can be calculated simply and quickly.

As shown in Figure 7, the first embodiment of the robot kinematic parameter self-calibration system of the present application includes: a calibration robot A and a reference robot B.

Among them, the flange of the calibration robot A is installed with an end effector 10, and the flange of the reference robot B is installed with a standard workpiece 20.

The reference robot B is used to change the pose of the standard workpiece 20.

The calibration robot A is used to execute instructions to implement the method provided in the first or second embodiment of the robot kinematic parameter self-calibration method of this application.

Among them, the number of calibrated kinematics parameters is related to the type of calibration robot A. In this embodiment, the calibration robot A can be a six-axis tandem robot.

In order to enable the calibration robot A to touch different positions on the surface 201 of the standard workpiece 20, the calibration robot A and the reference robot B are arranged relative to each other, so that the smart working spaces of the calibration robot A and the reference robot B have a larger overlap, that is, The moving space of the two robot arms has a large overlap, for example, the overlap reaches at least 60%.

In order to meet the requirements of calibration accuracy, the end effector 10 of the calibration robot A can use a machine tool probe with a repeatability less than or equal to 1um, and the standard workpiece 20 can use a three-coordinate inspection tool with higher sphericity and surface smoothness. Standard ball.

In the robot kinematics parameter self-calibration system of this embodiment, the reference robot with standard workpieces mounted on the flange is used to assist the calibration. The calibration tool is simple and efficient. During the calibration, only the movement of the calibration robot and the reference robot needs to be input once. In the process of calibration, the calibration robot and the reference robot can automatically move and calculate without manual intervention, so the calibration has a high degree of automation, which is suitable for large-scale calibration requirements in the robot production process.

As shown in FIG. 8, the structure of the second embodiment of the robot kinematic parameter self-calibration system of the present application is similar to the structure of the first embodiment of the robot kinematic parameter self-calibration system of the present application, except that it further includes: a control device C , Connect the calibration robot A and the reference robot B to control the movement of the calibration robot A and the reference robot B.

Wherein, the control device C can be a computer, a background server or a master controller and other control equipment, which can output control instructions to the calibration robot A and the reference robot B, and control the movement of the calibration robot A and the reference robot B.

In addition, the control device C can also store the movement data of the calibration robot A and the reference robot B, such as the angle value of each axis. The control device C can also receive a manipulation instruction from a user or a controller, so as to control the movement of the calibration robot A and the reference robot B according to the manipulation instruction.

As shown in FIG. 9, in an embodiment of the storage device of the present application, a program file 901 is stored inside the storage device 90, and the program file 901 can be executed to realize the first or second embodiment of the robot kinematic parameter self-calibration method of the present application The method provided.

Wherein, the storage device 90 can be a portable storage medium such as a U disk, an optical disk, or other storage devices such as a hard disk, etc., and can also be a server, a mobile terminal, a robot, or an independent component that can be integrated into the above device, such as a master Chip etc.

In this embodiment, when the program stored in the storage device is executed, a reference robot with a standard workpiece mounted on the flange is used to assist the calibration. The calibration tool is simple and efficient. During calibration, only one input of the calibration robot and the reference robot is required. In the calibration process, the calibration robot and the reference robot can automatically move and calculate without manual intervention, so the calibration has a high degree of automation, which is suitable for large-scale calibration requirements in the robot production process.

The above are only implementations of this application, and do not limit the scope of this application. Any equivalent structure or equivalent process transformation made by using the description and drawings of this application, or directly or indirectly applied to other related technologies In the same way, all fields are included in the scope of patent protection of this application.

Claims

A method for self-calibration of robot kinematics parameters, characterized in that it comprises:

The calibration robot touches the end effector to the surface of the standard workpiece multiple times, and records the angle of each axis of the calibration robot at each contact, and the standard workpiece is installed on the flange of the reference robot;

The reference robot changes the pose of the standard workpiece, and the calibration robot repeatedly executes the steps of contacting the surface of the standard workpiece with the end effector multiple times and recording the angles of each axis of the calibration robot during each contact to Acquiring multiple sets of axis angle values of the calibration robot;

Using the multiple sets of axis angle values, the nominal kinematic parameters of the calibration robot, and the actual radius of the standard workpiece, the actual kinematic parameters of the calibration robot are obtained.
The method according to claim 1, wherein the end effector of the calibration robot contacts the surface of the standard workpiece multiple times, and recording the angle of each axis of the calibration robot during each contact comprises:

The end effector of the calibration robot contacts different positions on the surface of the standard workpiece m times in different postures, and records the angle of each axis of the calibration robot at each contact.
The method according to claim 2, wherein the reference robot changes the pose of the standard workpiece, and the calibration robot repeatedly executes the end effector contacting the surface of the standard workpiece multiple times, and records each contact The step of calibrating the angles of each axis of the robot to obtain multiple sets of axis angle values of the calibration robot includes:

The reference robot changes the pose of the standard workpiece n-1 times;

Every time the posture of the standard workpiece is changed, the calibration robot will contact the end effector at different positions on the surface of the standard workpiece in different postures, and record the angle of each axis of the calibration robot during each contact to obtain m *The angle of each axis of the calibration robot described in group n.
The method according to claim 1, wherein the actual movement of the calibration robot is obtained by using the multiple sets of axis angle values, the nominal kinematic parameters of the calibration robot, and the actual radius of the standard workpiece Learning parameters include:

Using the multiple sets of axis angle values and the nominal kinematics parameters of the calibration robot, the theoretical position of the contact point between the end effector and the surface of the standard workpiece corresponding to each set of the axis angle values is calculated, wherein The angles of each axis of the calibration robot recorded under the same pose of the standard workpiece form a set of axis angle values;

Using multiple sets of the theoretical positions of the contact points, the fitting radius of the standard workpiece corresponding to each set of the theoretical positions of the contact points is obtained by fitting, wherein a plurality of the contact points are calculated by a set of the axis angle values The theoretical position is a set of theoretical positions of the contact points;

The actual kinematics parameters are calculated according to the fitted radius and the fitted radius error between the actual radius.
The method according to claim 4, wherein:

Before calculating the actual kinematic parameters according to the fitted radius and the fitted radius error between the actual radius, the method includes:

Establishing a first linear relationship equation between the position error of the contact point and the kinematic parameter error, wherein the kinematic parameter error is the error between the actual kinematic parameter and the nominal kinematic parameter;

Establishing a second linear relationship equation between the fitting radius error and the position error of the contact point;

Combining the first linear relationship equation and the second linear relationship equation to establish a third linear relationship equation between the fitting radius error and the kinematic parameter error;

The calculating the actual kinematics parameters according to the fitted radius and the fitted radius error between the actual radius includes:

The fitting radius error is substituted into the third linear relationship equation, and the third linear relationship equation is solved to obtain the actual kinematic parameters of the calibration robot.
The method according to claim 5, wherein the establishing the first linear relationship equation between the position error of the contact point and the kinematic parameter error comprises:

Using the difference between the theoretical position and the actual position of the contact point as an element to establish an error matrix of the position of the contact point;

Using the difference between the nominal kinematics parameter and the actual kinematics parameter as an element to establish an error matrix of the kinematics parameter;

Acquiring the first Jacobian matrix of the theoretical position of the contact point with respect to the kinematic parameter;

Establish the first linear relationship equation shown in the following formula:

Δp=J DH ΔDH;

Wherein, Δp represents the error matrix of the position of the contact point, J DH represents the first Jacobian matrix, and ΔDH represents the error matrix of the kinematic parameter.
The method according to claim 6, wherein the establishing a second linear relationship equation between the fitting radius error of the standard workpiece and the position error of the contact point comprises:

Using the difference between the fitting radius of the standard workpiece and the actual radius as an element to establish a fitting radius error matrix of the standard workpiece;

Acquiring a second Jacobian matrix of the fitting radius of the standard workpiece with respect to the theoretical position of the contact point;

Establish the second linear relationship equation shown in the following formula:

Δr=J p Δp;

Where, Δr represents the fitting radius error matrix of the standard workpiece, J p represents the second Jacobian matrix, and Δp represents the error matrix of the position of the contact point.
The method according to claim 7, wherein the combination of the first linear relationship equation and the second linear relationship equation establishes the difference between the fitting radius error of the standard workpiece and the kinematic parameter error The third linear relationship equation between the two includes:

Substituting the first linear relationship equation into the second linear relationship equation to obtain the third linear relationship equation shown in the following formula:

Δr=JΔDH;

Among them, J=J p J DH .
The method according to claim 5, wherein the solving the third linear relationship equation to obtain the actual kinematic parameters of the calibration robot comprises:

Solving the third linear relation equation to obtain the error matrix of the kinematic parameters of the calibration robot;

The nominal kinematics parameters of the calibration robot and the error matrix of the kinematics parameters are superimposed to obtain the actual kinematics parameters of the calibration robot.
The method according to any one of claims 1-9, wherein the calibration robot and the reference robot are arranged relative to each other.
The method according to claim 2, wherein the calibration robot is a six-axis tandem robot, and the value of m is greater than 4.
The method according to claim 3, wherein the calibration robot is a six-axis tandem robot, and the value of n is greater than 20.
The method according to claim 1, wherein the standard workpiece is a three-coordinate gauge standard ball.
A self-calibration system for robot kinematics parameters, which is characterized by comprising: a calibration robot and a reference robot;

An end effector is installed on the flange of the calibration robot, and a standard workpiece is installed on the flange of the reference robot;

The reference robot is used to change the pose of the standard workpiece;

The calibration robot is used to execute instructions to realize the robot kinematic parameter self-calibration method according to any one of claims 1-13.
The system according to claim 14, wherein the calibration robot and the reference robot are arranged oppositely.
The system according to claim 14, wherein the end effector of the calibration robot is a machine tool probe with a repeatability less than or equal to 1um.
The system according to claim 14, wherein the standard workpiece is a three-coordinate gauge standard ball.
The system according to claim 14, further comprising: a control device connected to the calibration robot and the reference robot for controlling the movement of the calibration robot and the reference robot.
The system according to claim 14, wherein the calibration robot is a six-axis tandem robot.
A storage device in which a program file is stored, wherein the program file is executed to realize the robot kinematic parameter self-calibration method according to any one of claims 1-13.