Spring constant correction device, spring constant correction method, and recording medium

Abstract

The spring constant correction device and method of the present invention, when a multi-joint robot is moving in a state where elastic deformation has been compensated, measures the position and posture of the distal end of the multi-joint robot, compares the measured values ​​of the distal end's position and posture with their target values, and corrects the spring constant based on the comparison result. During this correction, at a predetermined position, the spring constant is corrected based on at least three of the following: the angle of the endpoint posture obtained based on the measured posture, the torque of the distal end's endpoint posture, the angle of the distal end under the target posture, and the torque of the distal end under the target posture. The procedure for this method is recorded in the recording medium of the present invention.

Description

Technical Field

[0001] This invention relates to a spring constant correction device and method for use in multi-joint robots to correct the spring constant as a spring element, wherein the multi-joint robot allows a reducer to function as the spring element and undergo elastic deformation. The invention also relates to a recording medium containing the spring constant correction program. Background Technology

[0002] When welding multiple base materials using arc welding, weaving welding is often employed. This involves advancing the welding electrode along the welding direction while simultaneously performing a sinusoidal oscillating motion to the left and right of the weld line. This oscillating welding is typically achieved by swaying the welding torch left and right, or by tilting the torch left and right around its center. When performing this type of oscillating welding on a multi-joint robot, high trajectory accuracy is required. For example, the spring constant correction device for a multi-joint robot disclosed in Patent Document 1 achieves high trajectory accuracy while taking into account the elastic deformation of the reducer in the robot's power transmission system.

[0003] The spring constant correction device for a multi-joint robot disclosed in Patent Document 1 is applied to a multi-joint robot that uses a reducer as a spring element to perform elastic deformation. The multi-joint robot operates in a state where the elastic deformation is compensated for by an elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction device includes: a comparison unit that compares the measured values ​​of the position and posture of the distal end of the multi-joint robot, measured by sensors, when the multi-joint robot operates in the state where the elastic deformation has been compensated, with the target values ​​of the position and posture of the distal end of the multi-joint robot; and a correction unit that corrects the spring constant using the result obtained by the comparison unit. The correction unit converts the error of the position and posture of the distal end obtained by the comparison unit into a robot joint angle error, calculates the estimated torque at the swinging distal end position based on the model of the multi-joint robot, calculates the spring constant error based on the relationship between the robot joint angle error and the estimated torque, and uses the calculated spring constant error to correct the spring constant.

[0004] However, for the spring constant correction device for the multi-joint robot disclosed in Patent Document 1, the correction accuracy begins to decrease as the difference between the arm torque at the end position of the swing motion and the torque required at its target end position increases, so there is still room for improvement.

[0005] Existing technical documents

[0006] Patent documents

[0007] Patent Document 1: Japanese Invention Patent No. 6091272 (Published in Japanese Patent Publication No. 2014-180726) Summary of the Invention

[0008] This invention was made in view of the above circumstances, and its object is to provide a spring constant correction device and a spring constant correction method that can further improve the correction accuracy. Another object of this invention is to provide a recording medium on which a spring constant correction program is recorded.

[0009] The spring constant correction device and method of this invention involve measuring the position and posture of the distal end of a multi-joint robot when the robot is moving in a state where elastic deformation has been compensated. The measured values ​​of the distal end's position and posture are compared with target values, and the spring constant is corrected based on the comparison result. During correction, at a predetermined position, the spring constant is corrected based on at least three of the following: the angle of the endpoint posture obtained based on the measured distal end's posture; the torque of the distal end's endpoint posture; the angle of the distal end's target posture; and the torque of the distal end's target posture. The program of this spring constant correction method is recorded in a recording medium of this invention.

[0010] The above and other objects, features and advantages of the present invention will become clear from the following detailed description and drawings. Attached Figure Description

[0011] Figure 1 This is a schematic diagram illustrating the configuration of a multi-joint robot using a spring constant correction device in the implementation method.

[0012] Figure 2 This is a schematic diagram representing a model of one joint axis of the multi-joint robot.

[0013] Figure 3 This is a block diagram illustrating the configuration of the control device for the multi-joint robot, which includes the spring constant correction device.

[0014] Figure 4 This is a flowchart illustrating the swing inspection process of the multi-joint robot and the spring constant correction process performed by the spring constant correction device.

[0015] Figure 5 It is used for Figure 4The flowchart shown illustrates the process S5.

[0016] Figure 6 It is a diagram used to illustrate the relationship between the motor angle at the end point of the swing motion, the actual arm angle, the target arm angle, the actual arm torque, and the torque required for the target arm angle.

[0017] Figure 7 This is a diagram illustrating the relationship between the motor angle at the endpoint of the swing motion, the actual arm angle, the target arm angle, and the motor angle required to achieve the target arm angle under the corrected spring constant, in a modified embodiment. Detailed Implementation

[0018] Hereinafter, one or more embodiments of the present invention will be described with reference to the accompanying drawings. However, the scope of the invention is not limited to the disclosed embodiments. Furthermore, structures marked with the same symbols in the figures represent the same structures, and their descriptions are omitted as appropriate. In this specification, reference numerals without suffixes are used when referring to general terms, and reference numerals with suffixes are used when indicating individual structures.

[0019] The spring constant correction device involved in the embodiment is an apparatus applied to a multi-joint robot in which the reducer functions as a spring element and elastically deforms, and corrects the spring constant of the spring element. The multi-joint robot operates in a state where the elastic deformation is compensated for by an elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. For this type of multi-joint robot, the spring constant correction device includes: a measuring unit that measures the position and posture of the distal end of the multi-joint robot when the multi-joint robot operates in the state where the elastic deformation has been compensated; a comparison unit that compares the measured values ​​of the position and posture of the distal end of the multi-joint robot measured by the measuring unit with target values ​​of the position and posture of the distal end of the multi-joint robot; and a correction unit that corrects the spring constant based on the comparison result obtained by the comparison unit. Furthermore, the correction unit, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. Hereinafter, this spring constant correction device, as well as the spring constant correction method and procedure installed on the spring constant correction device, will be described in more detail.

[0020] Figure 1This is a schematic diagram illustrating the configuration of a multi-joint robot using a spring constant correction device in the implementation embodiment. Figure 2 This is a schematic diagram representing a model of one joint axis of the multi-joint robot. Figure 3 This is a block diagram illustrating the configuration of the control device for the multi-joint robot, which includes the spring constant correction device.

[0021] (Overall Composition)

[0022] The multi-joint robot 1 using the spring constant correction device in this embodiment is a so-called vertical multi-joint type, a six-axis robot with six degrees of freedom, consisting of six first joints JT1 to sixth joints JT6. Each of the first joints JT1 to sixth joints JT6 has an actuator that functions as a joint. In this embodiment, the actuator is, for example, a motor 15 (15-1 to 15-6). As an example, a welding torch is provided at the distal end of the axis of the sixth joint JT6 as a tool, and the multi-joint robot 1 can perform welding by arc welding using welding wire fed from the welding torch. The multi-joint robot 1 is configured such that the welding operation area is between a predetermined welding start point and a welding end point, and the multi-joint robot 1 moves along the welding line connecting the welding start point and the welding end point while performing an action (oscillating action) that tilts the welding wire with a predetermined amplitude and frequency. Furthermore, although the multi-joint robot using the spring constant correction device in this embodiment is a robot for welding purposes, it is not limited to this. For example, a suitable tool corresponding to the purpose of the multi-joint robot 1 can be set at the distal end of the axis of the sixth joint JT6 to replace the welding torch.

[0023] One joint axis of the multi-joint robot 1 can be as follows: Figure 2 The model is illustrated below. More specifically, the model representing a joint axis includes a motor 2, an encoder 3 installed on the motor 2 and serving as an example of an angle measuring unit for measuring the rotation angle of the motor 2, an arm 4 that rotates via the motor 2, and a reducer 5 connecting the motor 2 and the arm 4. The reducer 5 can be modeled as a spring element (elastic element). Therefore, when the spring constant is set to K, the rotation angle on the motor 2 side is set to θm, and the rotation angle on the arm 4 side is set to θl, the torque τl on the arm 4 side is represented by τl=K·(θl-θm).

[0024] (Control device: Robot controller as an example of such control device)

[0025] The multi-joint robot 1 is controlled by a robot controller 10, which is one example of the control device for the multi-joint robot 1. For example... Figure 3As shown, the robot controller 10 is connected to the measurement sensor 12 and the motors 15 (15-1 to 15-6) of each axis of the first joint JT1 to the sixth joint JT6 in the multi-joint robot 1. The measurement sensor 12 is a sensor that measures the position and posture of the distal end of the multi-joint robot 1. As described below, the multi-joint robot 1 moves in a state where elastic deformation has been compensated, so the measurement sensor 12 measures the position and posture of the distal end of the multi-joint robot 1 when the multi-joint robot 1 moves in the state where elastic deformation has been compensated. The measurement sensor 12 outputs its measurement results to the robot controller 10. The motors 15 of each axis of the first joint JT1 to the sixth joint JT6 in the multi-joint robot 1 are controlled by the robot controller 10 (e.g., servo control). In this control, because well-known and commonly used methods are used, therefore... Figure 3 The control block used for this control is not described in detail.

[0026] The robot controller 10 includes, for example, a computer and a teach pendant. The computer includes, for example, input devices, output devices, a memory, an interface device, a CPU (Central Processing Unit) that controls these devices and performs prescribed information processing, and their peripheral circuitry. The robot controller 10 stores various prescribed programs and data required for the jointed robot 1 to perform actions in its memory. Among these prescribed programs is, for example, a control processing program, which includes, for example, a control program, an elastic deformation compensation program, a target position comparison program, and a spring constant correction program. The control program controls the measurement sensor 12 to acquire measurement results and performs servo control on the motors 15 of each axis. The elastic deformation compensation program causes the jointed robot 1 to move in a state where elastic deformation is compensated based on the spring constant K of the spring element. The target position comparison program compares the measured values ​​of the distal position and posture of the jointed robot 1 measured by the measurement sensor 12 with the target values ​​of the distal position and posture of the jointed robot 1. The spring constant correction program corrects the spring constant based on the comparison results obtained by the comparison program. The various specified data include, for example, the spring constant K, the initial value of the spring constant K, and the teaching program. As an example, such specified programs and data can be read from a recording medium (e.g., a USB memory or CD-ROM) containing these specified programs and data and stored in the computer's memory.

[0027] The robot controller 10, by executing the control program, functionally constitutes an elastic deformation compensation unit 11, a target position comparison unit 13, and a spring constant correction unit 14.

[0028] The robot controller 10 controls the articulated robot 1 in such a way that the welding torch mounted on the articulated robot 1 swings and moves along the welding line according to a pre-taught teaching program. The teaching program can be pre-made, for example, using the teach pendant connected to the robot controller 10, or it can be pre-made using an offline teaching system utilizing the higher-level computer.

[0029] The spring constant correction device in this embodiment is implemented using the robot controller 10. However, while the spring constant correction device is implemented using the robot controller 10 in this embodiment, it is not limited to this. For example, the spring constant correction device may also be implemented using a higher-level computer connected to the robot controller 10.

[0030] The elastic deformation compensation unit 11 includes a target position comparison unit 13 and a spring constant correction unit 14. The target position comparison unit 13 compares the measured values ​​of the position and posture of the distal end (in this embodiment, the distal end of the welding torch) of the multi-joint robot 1 in, for example, the XY plane coordinate system, as measured by the measurement sensor 12, with the target values ​​of the position and posture. The spring constant correction unit 14 corrects the spring constant (in this embodiment, the spring constant of the modeled reducer 5) K based on the comparison result obtained by the target position comparison unit 13. The elastic deformation compensation unit 11 uses the spring constant K corrected by the spring constant correction unit 14 to compensate for elastic deformation. The robot controller 10 controls each motor 15 (15-1 to 15-6) of each axis in the multi-joint robot 1.

[0031] (Oscillation inspection and spring constant correction)

[0032] Figure 4 This is a flowchart illustrating the swing inspection process of the multi-joint robot and the spring constant correction process performed by the spring constant correction device. Figure 5 It is used for Figure 4 The flowchart shown illustrates the process S5. Figure 6 It is a diagram used to illustrate the relationship between the motor angle at the end point of the swing motion, the actual arm angle, the target arm angle, the actual arm torque, and the torque required for the target arm angle.

[0033] This type of movement is controlled by the robot controller 10. For example, a delivery inspection is performed on the jointed robot 1 before the product is delivered after manufacturing. One of these delivery inspections is a left-right swing check. In this left-right swing check, the spring constant K of the reducer 5 is corrected from its initial value.

[0034] In this left-right swing check, Figure 4The elastic deformation compensation unit 11 performs left-right swing check processing (S1). For example, the robot controller 10 is set to check mode, and servo control is manually implemented, causing the distal end of the arm of the multi-joint robot 1 to swing left and right. In this case, such as Figure 5 As shown, the coordinate data of the swing's distal end position (e.g., vertical direction (Z direction), horizontal direction (X direction)) is acquired and stored. Furthermore, one of the six axes is explained here.

[0035] Next, the elastic deformation compensation unit 11 determines whether the condition is acceptable (S2). For example, if the deviation (the difference between the target value and the measured value) is below a preset threshold (acceptance / acceptance determination threshold), the elastic deformation compensation unit 11 determines that the condition is acceptable; if the deviation exceeds the acceptance / acceptance determination threshold, the elastic deformation compensation unit 11 does not determine that the condition is acceptable (unacceptable). If the determination result is acceptable (Yes), the elastic deformation compensation unit 11 then performs process S3; on the other hand, if the condition is unacceptable (No), the elastic deformation compensation unit 11 then performs process S4 to perform spring constant correction processing.

[0036] In process S3, the elastic deformation compensation unit 11 performs a pass / fail determination process. This pass / fail determination process is, for example, determining the spring constant K and storing it in the memory, and causing the robot controller 10 to transition from the inspection mode to the normal mode.

[0037] In process S4, the elastic deformation compensation unit 11 acquires various robot parameters of the multi-joint robot 1. For example, it acquires the inertia term J, nonlinear term C, gravity term G, and spring constant K before correction for each axis of the multi-joint robot 1 stored in the robot controller 10 or the upper-level computer. Furthermore, since the spring constant K is rewritten in each of the repeated processes, in process S4, the spring constant K calculated and stored from the previous spring constant correction process or the initial value of the spring constant K is acquired.

[0038] Next, the elastic deformation compensation unit 11 calculates the error at the endpoint position of the swing motion (S5) using the target position comparison unit 13. For example... Figure 5 As shown, the difference between the ideal waveform representing the position of the distal end of the multi-joint robot 1 (represented by the dashed line) and the measured waveform measured by the measuring sensor 12 is taken as the error. For each point in the horizontal direction of the ideal waveform, its vertical value is a predetermined constant value. (Measured waveform) The error dX is represented by Equation 1 below, and by Equation 2 below. Furthermore, in the text, it is not possible to label it directly above "X". Therefore, for convenience, the label at the top is... The “X” is recorded as The same applies to the cases of "^", "·" or "··" directly above, which will be discussed later.

[0039] [Formula 1]

[0040] Formula 1;

[0041] [Formula 2]

[0042] Formula 2; dX=[dx, dy, dz, dα, dβ, dγ]

[0043] In Equation 2, x, y, and z are the coordinates of the position of the distal end of the arm in the multi-joint robot 1, and α, β, and γ are the attitude angles (Euler angles or roll, pitch, and yaw angles) of the distal end of the arm.

[0044] Here, while it would be ideal to measure all six degrees of freedom (x, y, z, α, β, γ), it is also possible to measure only at least two of the six degrees of freedom (x, y, z, α, β, γ) (for example, in the case of left-right oscillation, measuring the two degrees of freedom in the horizontal and vertical directions, which are the main directions of motion), assuming dα, dβ, and dγ as 0, and dx or dy as 0. Thus, even when two sensors are used to measure the position (or posture) of the distal end of the arm in the multi-joint robot 1, the spring constant correction method of this embodiment can be implemented. Furthermore, the spring constant correction method is not limited to the assumptions described above.

[0045] Next, the elastic deformation compensation unit 11, as shown in Equation 3 below, uses the Jacobian matrix to convert the position error into the angle error dθ of the joint JT (S6).

[0046] [Formula 3]

[0047] Equation 3; dθ=J(θ0) -1 ·dX

[0048] (θo: Angles of each axis at the center of oscillation)

[0049] When correcting (updating) the spring constant K, firstly, the Jacobian matrix is ​​used to convert the error of the position and posture of the far end of the multi-joint robot 1, which is the result of the comparison obtained by the target position comparison unit 13, into the angle error dθ of the joint JT.

[0050] Next, the elastic deformation compensation unit 11 corrects (updates) the spring constant K through the spring constant correction unit 14, and returns the process to process S1 (S7). More specifically, the spring constant K is corrected in the following manner.

[0051] For the axis of a joint JT, the motor angle θ at the endpoint of the swing motion. m The actual arm angle θ l Arm target angle θ d Actual arm torque τ l Arm target angle θ d Required torque τ d Physical relationships such as Figure 6 As shown. Furthermore, although each joint JT undergoes rotational movement, but in Figure 6 For simplicity, the rotation angle is illustrated by converting it into the linear distance of linear motion springs. (The following will be discussed...) Figure 7 The same.

[0052] When controlling motor 15(2), robot controller 10 uses the current spring constant K. old The decision is made to make the arm angle reach the target arm angle θ. d Motor angle θ m These relationships are given by Equation 4 below.

[0053] [Formula 4]

[0054] Equation 4; τ d =K old (θ d -θ m )

[0055] On the other hand, based on the motion equation of the multi-joint robot 1 shown in Equation 5 below, the target angle θ of the arm is obtained. d Required torque τ d Here, τ d ^ is the torque represented by a vector with the torque of each axis as components, τ d It is τ d A component of ^. θ D θ D · and θ D ·· These are the angles, angular velocities, and angular accelerations of each axis under the target posture at the endpoint of the swing motion.

[0056] [Formula 5]

[0057] Formula 5;

[0058] θ D , : Angle, angular velocity, and angular acceleration of each axis under the target posture at the swing endpoint

[0059] If we set the corrected (updated true value) spring constant as Knew Because of the current spring constant K old Relative to the corrected spring constant K new Includes error d θ Therefore, the actual arm angle θ l Angle θ with the target arm d An error dθ will be generated between them. d =θ l -θ d Error dθ d It is a component of the vector error dθ. If the angle of motor 15(2) is affected by the motor angle θ m If controlled, these relationships are given by Equation 6 below.

[0060] [Formula 6]

[0061] Equation 6; τ l =K new (θ l -θ m )

[0062] On the other hand, the actual arm torque τ is obtained from the motion equation of the multi-joint robot 1 shown in Equation 7 below. l Here, τ l ^ is the torque represented by a vector with the torque of each axis as components, τ l It is τ l One of the components in ^. and These are the endpoints of the swinging motion. The angles, angular velocities, and angular accelerations of each axis at each location.

[0063] [Formula 7]

[0064] Formula 7;

[0065] ( Swing End (Angle, angular velocity, angular acceleration of each axis at the location)

[0066] Based on Equations 4 and 6, we obtain Equation 8.

[0067] [Formula 8]

[0068] Formula 8;

[0069] All the variables on the right-hand side of Equation 8 are known or can be calculated from known data. Therefore, by using Equation 8, the current spring constant K can be determined. old Corrected (updated) to the corrected spring constant K new .

[0070] That is, the spring constant correction unit 14, at a predetermined position, determines the angle (θ) of the endpoint posture based on the posture of the distal end of the multi-joint robot measured by the measurement unit 12. l The torque (τ) of the distal end pose of the multi-joint robot. l The angle (θ) of the target pose at the distal end of the multi-joint robot. d ), and the torque (τ) at the distal end of the multi-joint robot under the target posture. d At least three of the following are used to correct the spring constant K. More specifically, in this embodiment, the corrected spring constant K... new Represented by function 8, which uses the torque (τ) of the endpoint posture at the distal end of the multi-joint robot at the endpoint position of the swing motion. l The angle (θ) of the target pose at the distal end of the multi-joint robot. d ), and the torque (τ) at the distal end of the multi-joint robot under the target posture. d Although in the above example the specified position is the endpoint of the swinging motion, it could also be any other position within the swinging motion.

[0071] The correction process S7 for the spring constant K is implemented using Equation 8 as described above.

[0072] Alternatively, it could be the corrected spring constant K. new The following functional expression is used at the specified position, where the function uses the angle (θ) of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot 1 measured by the measuring unit 12. l The torque (τ) of the distal end pose of the multi-joint robot 1 l The angle (θ) of the target pose at the distal end of the multi-joint robot 1. d ), and the torque (τ) at the distal end of the multi-joint robot 1 in the target posture. d At least three of the spring constants are used, and the spring constant correction unit 14 uses the least squares method for the function to correct the spring constant. Because this spring constant correction device uses the least squares method, the function can be customized according to the multi-joint robot 1 to which the spring constant correction device is applied, so that the spring constant can be corrected with the correction accuracy most suitable for the multi-joint robot 1.

[0073] For example, by having the multi-joint robot 1 perform multiple swinging movements, collecting multiple data points, and using these collected data points in Equation 8, the corrected spring constant K can be obtained. newAs the least squares solution. Here, the corrected spring constant K can also be obtained by fitting multiple data to Equation 9, which is derived from Equation 8. new As a least squares solution. The least squares method can also use the weighted least squares method.

[0074] [Formula 9]

[0075] Equation 9;

[0076] As explained above, the spring constant correction device, the spring constant correction method, and the spring constant correction procedure installed on the spring constant correction device in the embodiment, at a predetermined position (in the above example, the endpoint position of the swinging motion), are based on the angle (θ) of the endpoint posture obtained from the posture of the distal end of the multi-joint robot 1 measured by the measuring unit 12. l The torque (τ) of the distal end pose of the multi-joint robot 1 l The angle (θ) of the target pose at the distal end of the multi-joint robot 1. d ), and the torque (τ) at the distal end of the multi-joint robot 1 in the target posture. d The spring constant is corrected by using at least three of the following methods: (1) the spring constant is calibrated by (2) the spring constant is calibrated by (3) the spring constant is calibrated by (4) the spring constant is calibrated by (5) the spring constant is calibrated by (6) the spring constant is calibrated by (7) the spring constant is calibrated by (8) the spring constant is calibrated by (9) the spring constant is calibrated by (10) the spring constant is calibrated by (11) the spring constant is calibrated by (2) the spring constant is calibrated by (3) the spring constant is calibrated by (4) the spring constant is calibrated by (5) the spring constant is calibrated by (6) the spring constant is calibrated by (7) the spring constant is calibrated by (8) the spring constant is calibrated by (9) the spring constant is calibrated by (10) the spring constant is

[0077] In the aforementioned patent document 1, the corrected spring constant K is calculated according to the following formula 10. new (Section

[0032] of the aforementioned Patent Document 1). It is now understood that, if Equation 8 is modified according to the viewpoint of Equation 10, it can be modified in the manner of Equation 11, and it is now understood that Equation 10 corresponds to τ in Equation 11. l =τ d This understanding leads to the conclusion that, for the spring constant correction device of the multi-joint robot disclosed in Patent Document 1, the correction accuracy will vary with the arm torque τ at the end position of the swing motion. l With the torque τ required at the target endpoint position d The difference between them increases and then begins to decrease, which is consistent with the experience so far. The spring constant correction device of this embodiment, as described above, adjusts the arm torque τ at the end position of the swing motion regardless of the position of the swing. l With the torque τ required at the target endpoint position d Regardless of the differences between them, the spring constant can be corrected, thus improving the reduction in the correction accuracy.

[0078] [Formula 10]

[0079] Formula 10;

[0080] K new Updated spring constant

[0081] K old Spring constant before update

[0082] [Formula 11]

[0083] Formula 11;

[0084] According to this embodiment, a spring constant correction device, a spring constant correction method, and a spring constant correction program using function 8 can be provided.

[0085] Furthermore, in the above embodiment, the spring constant K is represented by function 8 (first function), but it is not limited to this. The function 8 (first function) uses the torque (τ) of the distal end posture of the multi-joint robot at the endpoint position of the swinging motion. l The angle (θ) of the target pose at the distal end of the multi-joint robot. d ), and the torque (τ) at the distal end of the multi-joint robot under the target posture. d ).

[0086] For example, the spring constant K can also be represented by function 15 (second function) described later, which uses the angle (θ) of the endpoint posture obtained at the endpoint position of the swinging motion based on the posture of the distal end of the multi-joint robot measured by the measuring unit. l The angle (θ) of the target pose at the distal end of the multi-joint robot. d ), and the torque (τ) at the target pose of the distal end of the multi-joint robot. d ).

[0087] Alternatively, for example, the spring constant K can also be represented by function 16 (third function) described later, which uses the angle (θ) of the endpoint posture obtained at the endpoint position of the swinging motion based on the posture of the distal end of the multi-joint robot measured by the measuring unit 12. l The torque (τ) of the distal end pose of the multi-joint robot. l ), and the angle (θ) of the target pose at the distal end of the multi-joint robot. d ).

[0088] Figure 7This is a diagram illustrating the relationship between the motor angle at the endpoint of the swing motion, the actual arm angle, the target arm angle, and the motor angle required to achieve the target arm angle, in a modified embodiment.

[0089] In this modified embodiment, such as Figure 7 As shown, for the corrected spring constant K new In this case, consider making the arm angle reach the target arm angle θ. d Required motor angle θ x These relationships are given by Equation 12 below.

[0090] [Formula 12]

[0091] Equation 12; τ d =K new (θ d -θ X )

[0092] Here, when the motor 15(2) and the arm (4) are connected by a linear spring, the ratio of the amplitude of the motor angle to the amplitude of the arm angle depends only on the vibration frequency (oscillation frequency). Therefore, for the same frequency of action, Equation 13 holds.

[0093] [Formula 13]

[0094] Equation 13;

[0095] Based on equations 6, 12, and 13, equation 14 is obtained. By using equation 14, τ is eliminated from equation 8. l Equation 15 is obtained, and τ is eliminated from Equation 8 by using Equation 14. d , thus obtaining equation 16.

[0096] [Formula 14]

[0097] Equation 14;

[0098] [Formula 15]

[0099] Formula 15;

[0100] [Formula 16]

[0101] Formula 16;

[0102] Furthermore, in the above embodiment, although the torque (τ) of the endpoint posture can be calculated from the motion equation of the multi-joint robot 1, lHowever, it may also include a second measuring unit (second measuring sensor) that measures the torque of the distal end-point posture of the articulated robot 1. Since the articulated robot 1 operates in a state where elastic deformation has been compensated, this second measuring sensor measures the torque of the distal end-point posture of the articulated robot 1 when the articulated robot 1 operates in this compensated elastic deformation state. Furthermore, the second measuring sensor outputs the measurement result to the robot controller 10.

[0103] As described above, this specification discloses various implementation techniques, the main techniques of which are summarized below.

[0104] One aspect of the spring constant correction device involves a multi-joint robot that elastically deforms due to the reducer acting as a spring element. The multi-joint robot operates in a state where elastic deformation is compensated for by an elastic deformation compensation unit included in a robot controller based on the spring constant of the spring element. The spring constant correction device includes: a measuring unit that measures the position and posture of the distal end of the multi-joint robot when the robot operates in the state where the elastic deformation has been compensated; a comparison unit that compares the measured values ​​of the position and posture of the distal end of the multi-joint robot measured by the measuring unit with target values ​​of the position and posture of the distal end of the multi-joint robot; and a correction unit that corrects the spring constant based on the comparison result obtained by the comparison unit. The correction unit, at a predetermined position, corrects the spring constant based on at least three of the following: an angle of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot measured by the measuring unit; a torque of the endpoint posture of the distal end of the multi-joint robot; an angle of the distal end of the multi-joint robot in the target posture; and a torque of the distal end of the multi-joint robot in the target posture. Ideally, in the aforementioned spring constant correction device, the spring constant correction unit calculates the torque of the endpoint posture based on the motion equation of the multi-joint robot. Ideally, the aforementioned spring constant correction device also includes a second measuring unit that measures the torque of the endpoint posture at the distal end of the multi-joint robot when the multi-joint robot performs an action in a state where the elastic deformation has been compensated.

[0105] This type of spring constant correction device, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. Therefore, the spring constant can be corrected regardless of the difference between the arm torque at the endpoint position of the swinging motion and the torque required at the target endpoint position. Thus, the above-described spring constant correction device can further improve the correction accuracy.

[0106] In another embodiment, in the above-described spring constant correction device, the spring constant is expressed as a function. At the specified position, the function uses at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit, the torque of the endpoint posture of the distal end of the multi-joint robot, the angle of the distal end of the multi-joint robot under the target posture, and the torque of the distal end of the multi-joint robot under the target posture. The correction unit uses the least squares method for the function to correct the spring constant.

[0107] Because this spring constant correction device uses the least squares method, the function can be customized according to the joint robot to which the spring constant correction device is applied, so that the spring constant can be corrected with the most suitable correction accuracy for the joint robot.

[0108] In another embodiment, in these above-described spring constant correction devices, the specified position is the endpoint position of the swinging motion, and the spring constant is expressed by a function. At the specified position, the function uses the torque of the endpoint posture of the distal end of the multi-joint robot, the angle of the distal end of the multi-joint robot in the target posture, and the torque of the distal end of the multi-joint robot in the target posture. Ideally, in the above-described spring constant correction device, the function is Equation 8. Thus, a spring constant correction device using a function can be provided in one embodiment.

[0109] In another embodiment, in these spring constant correction devices described above, the predetermined position is the endpoint position of the swinging motion, and the spring constant is expressed by a function. At the predetermined position, the function uses the angle of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot measured by the measuring unit, the angle of the distal end of the multi-joint robot under the target posture, and the torque of the distal end of the multi-joint robot under the target posture. Ideally, in the above-described spring constant correction device, the function is Equation 15. Thus, another embodiment of a spring constant correction device using a function can be provided.

[0110] In another embodiment, in these spring constant correction devices described above, the predetermined position is the endpoint position of the swinging motion, and the spring constant is expressed by a function. At the predetermined position, the function uses the angle of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot measured by the measuring unit, the torque of the endpoint posture of the distal end of the multi-joint robot, and the angle of the target posture of the distal end of the multi-joint robot. Ideally, in the above-described spring constant correction device, the function is Equation 16. Thus, another embodiment of a spring constant correction device using a function can be provided.

[0111] Another embodiment involves a spring constant correction method applied to a multi-joint robot that elastically deforms due to the reducer acting as a spring element. The multi-joint robot operates in a state where elastic deformation is compensated for by an elastic deformation compensation unit included in a robot controller based on the spring constant of the spring element. The spring constant correction method includes: a measurement step, measuring the position and posture of the distal end of the multi-joint robot when it operates in the state where the elastic deformation has been compensated; a comparison step, comparing the measured values ​​of the position and posture of the distal end of the multi-joint robot obtained by the first measurement step with target values ​​of the position and posture of the distal end of the multi-joint robot; and a correction step, correcting the spring constant based on the result obtained by the comparison step. The correction step, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot measured by the measurement step; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot under the target posture; and the torque of the distal end of the multi-joint robot under the target posture.

[0112] Another embodiment involves a spring constant correction procedure applied to a multi-joint robot that elastically deforms as a spring element, and executed by a computer. The multi-joint robot operates in a state where the elastic deformation is compensated for by an elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction procedure includes: a measurement step, measuring the position and posture of the distal end of the multi-joint robot when it operates in the state where the elastic deformation has been compensated; a comparison step, comparing the measured values ​​of the position and posture of the distal end of the multi-joint robot obtained by the first measurement step with target values ​​of the position and posture of the distal end of the multi-joint robot; and a correction step, correcting the spring constant based on the result obtained by the comparison step. The correction step, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained based on the posture of the distal end of the multi-joint robot measured by the measurement step; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture.

[0113] These spring constant correction methods and procedures, at a specified position, correct the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. Therefore, the spring constant can be corrected regardless of the difference between the arm torque at the endpoint position of the swinging motion and the torque required at the target endpoint position. Thus, the aforementioned spring constant correction methods and procedures can further improve correction accuracy.

[0114] Another embodiment involves a recording medium that records the aforementioned spring constant correction procedure. This allows for the provision of a recording medium that records the aforementioned spring constant correction procedure.

[0115] This application is based on Japanese Patent Application No. 2021-024925, filed on February 19, 2021, the contents of which are included in this application.

[0116] To illustrate the present invention, the invention has been appropriately and sufficiently described above in conjunction with the accompanying drawings and embodiments. However, it should be understood that those skilled in the art can readily modify and / or improve the above embodiments. Therefore, it should be interpreted that any modifications or improvements implemented by those skilled in the art that do not depart from the scope of the claims are still included within the scope of the claims.

[0117] Industrial applicability

[0118] According to the present invention, a spring constant correction device, a spring constant correction method, and a recording medium can be provided.

Claims

1. A spring constant correction device, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction device includes: The measurement unit measures the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison unit compares the measured values ​​of the position and posture of the distal end of the multi-joint robot, as measured by the measurement unit, with the target values ​​of the position and posture of the distal end of the multi-joint robot; and, The correction unit, based on the comparison result obtained by the comparison unit, corrects the spring constant, wherein... The correction unit, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is represented by the following function 1: At the specified position, function 1 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measuring unit. l The torque τ of the distal end pose of the multi-joint robot l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

2. The spring constant correction device according to claim 1, characterized in that: The correction unit uses the least squares method for the function 1 to correct the spring constant.

3. A spring constant correction device, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction device includes: The measurement unit measures the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison unit compares the measured values ​​of the position and posture of the distal end of the multi-joint robot, as measured by the measurement unit, with the target values ​​of the position and posture of the distal end of the multi-joint robot; and, The correction unit, based on the comparison result obtained by the comparison unit, corrects the spring constant, wherein... The correction unit, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 2: At the specified position, function 2 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measuring unit. l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

4. The spring constant correction device according to claim 3, characterized in that: The correction unit uses the least squares method for the function 2 to correct the spring constant.

5. A spring constant correction device, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction device includes: The measurement unit measures the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison unit compares the measured values ​​of the position and posture of the distal end of the multi-joint robot, as measured by the measurement unit, with the target values ​​of the position and posture of the distal end of the multi-joint robot; and, The correction unit, based on the comparison result obtained by the comparison unit, corrects the spring constant, wherein... The correction unit, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured by the measuring unit; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 3: At the specified position, function 3 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measuring unit. l The torque τ of the distal end pose of the multi-joint robot l and the angle θ of the target pose at the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

6. The spring constant correction device according to claim 5, characterized in that: The correction unit uses the least squares method for the function 3 to correct the spring constant.

7. A spring constant correction method, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction method includes the following steps: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison process involves comparing the measured values ​​of the position and posture of the distal end of the multi-joint robot obtained in the measurement process with the target values ​​of the position and posture of the distal end of the multi-joint robot. as well as, The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is represented by the following function 1: At the specified position, function 1 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The torque τ of the distal end pose of the multi-joint robot l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

8. A recording medium applied to a multi-joint robot in which a reducer functions as a spring element and elastically deforms, wherein a spring constant correction program executed by a computer is recorded, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction procedure includes: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison step involves comparing the measured values ​​of the distal position and orientation of the multi-joint robot obtained in the measurement step with the target values ​​of the distal position and orientation of the multi-joint robot; and... The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is represented by the following function 1: At the specified position, function 1 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The torque τ of the distal end pose of the multi-joint robot l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d For angular error, The recording medium contains the spring constant correction program and can be read by a computer.

9. A spring constant correction method, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction method includes the following steps: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison process involves comparing the measured values ​​of the position and posture of the distal end of the multi-joint robot obtained in the measurement process with the target values ​​of the position and posture of the distal end of the multi-joint robot. as well as, The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 2: At the specified position, function 2 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

10. A recording medium applied to a multi-joint robot in which a reducer functions as a spring element and elastically deforms, wherein a spring constant correction program executed by a computer is recorded, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction procedure includes: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison step involves comparing the measured values ​​of the distal position and orientation of the multi-joint robot obtained in the measurement step with the target values ​​of the distal position and orientation of the multi-joint robot; and... The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 2: At the specified position, function 2 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The angle θ of the target pose at the distal end of the multi-joint robot. d and the torque τ at the target posture of the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d For angular error, The recording medium contains the spring constant correction program and can be read by a computer.

11. A spring constant correction method, applied to a multi-joint robot in which the reducer functions as a spring element and undergoes elastic deformation, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction method includes the following steps: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison process involves comparing the measured values ​​of the position and posture of the distal end of the multi-joint robot obtained in the measurement process with the target values ​​of the position and posture of the distal end of the multi-joint robot. as well as, The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 3: At the specified position, function 3 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The torque τ of the distal end pose of the multi-joint robot l and the angle θ of the target pose at the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d This represents the angular error.

12. A recording medium applied to a multi-joint robot in which a reducer functions as a spring element and elastically deforms, recording a spring constant correction program executed by a computer, characterized in that: The multi-joint robot operates in a state where the elastic deformation is compensated for by the elastic deformation compensation unit included in the robot controller based on the spring constant of the spring element. The spring constant correction procedure includes: The measurement process involves measuring the position and posture of the distal end of the multi-joint robot when the elastic deformation has been compensated. The comparison step involves comparing the measured values ​​of the distal position and orientation of the multi-joint robot obtained in the measurement step with the target values ​​of the distal position and orientation of the multi-joint robot; and... The calibration process, based on the results of the comparison obtained in the comparison process, corrects the spring constant, wherein... The calibration process, at a predetermined position, corrects the spring constant based on at least three of the following: the angle of the endpoint posture obtained from the posture of the distal end of the multi-joint robot measured in the measurement process; the torque of the endpoint posture of the distal end of the multi-joint robot; the angle of the distal end of the multi-joint robot in the target posture; and the torque of the distal end of the multi-joint robot in the target posture. The specified position is the endpoint of the swinging motion. The spring constant is expressed by the following function 3: At the specified position, function 3 uses an angle θ based on the end-point pose obtained from the pose of the distal end of the multi-joint robot measured by the measurement process. l The torque τ of the distal end pose of the multi-joint robot l and the angle θ of the target pose at the distal end of the multi-joint robot. d , Among them, K new K is the corrected spring constant. old Let dθ be the current spring constant. d For angular error, The recording medium contains the spring constant correction program and can be read by a computer.

Images