Apparatus, control device and method for generating signals for a weaving operation - Patents.com

The device generates and corrects weaving signals to stabilize robot operations, enhancing welding accuracy by suppressing resonance during tool oscillation.

JP7583176B2Active Publication Date: 2024-11-13FANUC LTD

Patent Information

Application Number
JP2023539428
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2021-08-03
Publication Date
2024-11-13
Estimated Expiration
2041-08-03

AI Technical Summary

Technical Problem

Robots performing weaving operations can resonate during tool oscillation, leading to a reduction in work quality.

Method used

A device and method that generates a weaving signal for a robot to move a tool along a predetermined path while swinging it in an intersecting direction, includes a resonance determination unit to detect resonance, and corrects the signal frequency to suppress resonance.

Benefits of technology

The weaving operation is performed more stably, improving the accuracy and quality of welding work by preventing resonance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007583176000001
    Figure 0007583176000001
  • Figure 0007583176000002
    Figure 0007583176000002
  • Figure 0007583176000003
    Figure 0007583176000003
Patent Text Reader

Abstract

There have been situations in which, when a robot swings a tool in a weaving motion, the robot would resonate and, consequently, work quality would suffer. A device 80 comprises: a weaving signal generation unit 56 that generates a weaving signal for causing a robot 12 to swing a tool; a chronological data acquisition unit 60 for acquiring chronological data of the amplitude value of the tool when the robot 12 is caused to execute a weaving motion according to the weaving signal; a frequency characteristic acquisition unit 62 for acquiring a first frequency characteristic of chronological data; a resonance determination unit 64 for determining, on the basis of the first frequency characteristic, whether or not the robot is resonating at the frequency of the weaving signal generated by the weaving signal generation unit; and a correction unit 68 for correcting the weaving signal so as to change frequencies if the robot 12 is determined to be resonating.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to an apparatus for generating a signal for a weaving operation. , regulation Control device and method Regarding. [Background technology]

[0002] A device that causes a robot to perform a weaving operation is known (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 10-235582 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventionally, when a robot swings a tool during a weaving operation, the robot may resonate, which may result in a decrease in work quality. [Means for solving the problem]

[0005] In one aspect of the present disclosure, an apparatus for generating a signal for a weaving operation in which a robot moves a tool along a predetermined work path and swings the tool in a direction intersecting the work path includes a weaving signal generation unit that generates a weaving signal for causing the robot to swing the tool, a time series data acquisition unit that acquires time series data of amplitude values ​​of the tool in the intersecting direction when the robot is made to perform a weaving operation in accordance with the weaving signal generated by the weaving signal generation unit, a frequency characteristic acquisition unit that acquires a first frequency characteristic of the time series data acquired by the time series data acquisition unit, a resonance determination unit that determines whether the robot is resonating at the frequency of the weaving signal generated by the weaving signal generation unit based on the first frequency characteristic acquired by the frequency characteristic acquisition unit, and a correction unit that corrects the weaving signal to change the frequency when it is determined by the resonance determination unit that the robot is resonating.

[0006] In another aspect of the present disclosure, a method for generating a signal for a weaving operation in which a robot moves a tool along a predetermined work path and oscillates the tool in a direction intersecting the work path includes a processor generating a weaving signal for causing the robot to oscillate the tool, acquiring time series data of amplitude values ​​of the tool in the intersecting direction when the robot is caused to perform a weaving operation in accordance with the generated weaving signal, acquiring a first frequency characteristic of the acquired time series data, determining whether the robot is resonating at the frequency of the generated weaving signal based on the acquired first frequency characteristic, and correcting the weaving signal to change the frequency when it is determined that the robot is resonating. Effect of the Invention

[0007] Since it is possible to suppress the resonance that occurs when the tool is swung during the weaving operation, it is possible to perform the weaving operation more stably, which in turn increases the accuracy of the welding work and thereby improves the welding quality. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram of a robotic system according to one embodiment. [Diagram 2] FIG. 2 is a block diagram of the robot system shown in FIG. [Diagram 3] 2 shows a trajectory of a weaving motion performed by the robot shown in FIG. 1. [Figure 4] 2 is a block diagram of a function for generating a weaving operation command in the control device shown in FIG. 1. FIG. [Diagram 5] 4 shows a waveform of a first weaving signal according to an embodiment. [Figure 6] 6 shows a waveform of a second weaving signal generated by filtering the first weaving signal shown in FIG. 5. [Figure 7] 7 shows frequency characteristics (frequency spectrum) of the second weaving signal shown in FIG. 6. [Figure 8] 4 shows the time change characteristics of the tool amplitude value when the robot resonates at the fundamental frequency. [Figure 9] 9 shows the frequency characteristics of the time-varying characteristics shown in FIG. 8. [Figure 10] This shows the time change characteristics of the tool amplitude value when the robot resonates at a harmonic frequency. [Figure 11] 9 shows the frequency characteristics of the time-varying characteristics shown in FIG. 8. [Figure 12] 2 is a flowchart showing an example of a control flow of a welding operation executed by the robot system shown in FIG. 1. [Figure 13] FIG. 13 is a diagram of a robot system according to another embodiment. [Figure 14] FIG. 14 is a block diagram of the robot system shown in FIG. 13. [Figure 15] FIG. 14 is a functional block diagram of the teaching device shown in FIG. 13 for generating a weaving operation command. [Figure 16] 4 shows another example of a trajectory of a weaving motion. [Figure 17] 17 shows a waveform of a weaving signal for causing the robot to perform a weaving operation along the trajectory shown in FIG. 16. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present disclosure will be described in detail with reference to the drawings. In various embodiments described below, similar elements are given the same reference numerals, and duplicated explanations will be omitted. First, a robot system 10 according to one embodiment will be described with reference to Figs. 1 and 2. The robot system 10 includes a robot 12 and a control device 14.

[0010] In this embodiment, the robot 12 performs a welding operation on a workpiece W. Specifically, the robot 12 is a vertical articulated robot, and has a robot base 16, a rotating body 18, a lower arm 20, an upper arm 22, a wrist 24, and a tool 26. The robot base 16 is fixed to the floor of a work cell or on an automated guided vehicle (AGV). The rotating body 18 is provided on the robot base 16 so as to be rotatable around a vertical axis.

[0011] The lower arm 20 is provided on the rotating body 18 so as to be rotatable around a horizontal axis. The upper arm 22 is rotatably provided at the tip of the lower arm 20. The wrist 24 has a wrist base 24a provided at the front end of the upper arm 22 so as to be rotatable around two axes that are perpendicular to each other, and a wrist flange 24b provided rotatably on the wrist base 24a.

[0012] Tool 26 is detachably attached to wrist flange 24b. In this embodiment, tool 26 is a welding torch that generates an electric discharge between itself and workpiece W in response to a command from control device 14, and melts a welding wire fed from a wire feeder (not shown) to weld workpiece W. Tool 26 may be a laser processing head that emits a laser beam to melt the welding wire with the laser beam and weld the workpiece W.

[0013] A plurality of servo motors 28 (FIG. 2) are provided on each of the robot base 16, the rotating body 18, the lower arm 20, the upper arm 22, and the wrist 24. These servo motors 28 rotate each movable element of the robot 12 (i.e., the rotating body 18, the lower arm 20, the upper arm 22, the wrist 24, and the wrist flange 24b) in response to a command from the control device 14, thereby moving the tool 26.

[0014] The control device 14 controls the operation of the robot 12. As shown in Fig. 2, the control device 14 is a computer having a processor 30, a memory 32, and an I / O interface 34. The processor 30 has a CPU or a GPU, etc., and is communicatively connected to the memory 32 and the I / O interface 34 via a bus 36, and performs calculation processing for executing a weaving operation described below while communicating with these components.

[0015] The memory 32 has a RAM, a ROM, or the like, and temporarily or permanently stores various data used in the arithmetic processing executed by the processor 30 and various data generated during the arithmetic processing. The I / O interface 34 has, for example, an Ethernet (registered trademark) port, a USB port, an optical fiber connector, or an HDMI (registered trademark) terminal, and communicates data with an external device via wired or wireless communication under instructions from the processor 30. In this embodiment, the servo motor 28 of the robot 12 is communicatively connected to the I / O interface 34.

[0016] The control device 14 is provided with an input device 38 and a display device 40. The input device 38 has a keyboard, a mouse, a touch panel, or the like, and receives data input from an operator. The display device 40 has a liquid crystal display, an organic EL display, or the like, and displays various data.

[0017] The input device 38 and the display device 40 may be communicatively connected to the I / O interface 34 via a wire or wirelessly. The input device 38 and the display device 40 may be provided separately from the housing of the control device 14, or may be integrated into the housing of the control device 14.

[0018] 1, a robot coordinate system C1 and a tool coordinate system C2 are set for the robot 12. The robot coordinate system C1 is a coordinate system for automatically controlling each movable element of the robot 12. In this embodiment, the robot coordinate system C1 is set for the robot 12 so that its origin is located at the center of the robot base 16 and its z axis coincides with the rotation axis of the rotating body 18. In the following description, for convenience, the positive x-axis direction of the robot coordinate system C1 will be referred to as the right, the positive y-axis direction as the front, and the positive z-axis direction as the up.

[0019] The tool coordinate system C2 is a coordinate system that defines the position of the tool 26 in the robot coordinate system C1, and is set with respect to the tool 26. In this document, "position" may mean position and attitude. In this embodiment, the tool coordinate system C2 is set with respect to the tool 26 so that its origin (so-called TCP) is located at the working point of the tool 26 (e.g., the point where the tool 26 melts the welding wire) and its z-axis direction coincides with the working direction of the tool 26 (e.g., the direction of the central axis of the tip of the welding torch, or the emission direction of the laser light).

[0020] When moving the tool 26, the processor 30 sets a tool coordinate system C2 in the robot coordinate system C1, and sends commands to the servo motors 28 of the robot 12 to operate the movable elements of the robot 12 so as to position the tool 26 at a position represented by the set tool coordinate system C2. In this way, the processor 30 positions the tool 26 at any position in the robot coordinate system C1 by the operation of the robot 12.

[0021] Next, with reference to Fig. 3, a weaving operation that the processor 30 causes the robot 12 to execute during welding work on the workpiece W will be described. The weaving operation is executed to widen the width of a bead formed on the workpiece W when the workpiece W is welded. Specifically, in the weaving operation, the processor 30 causes the robot 12 to move the tool 26 along a predetermined work path WP and swing the tool 26 in a direction intersecting the work path WP.

[0022] 3, the working path WP is set on the workpiece W so as to extend linearly in the x-axis direction of the robot coordinate system C1. In this example, in the weaving operation, the processor 30 causes the robot 12 to move the tool 26 to the right along the working path WP and swing the tool 26 back and forth. As a result, the tool 26 welds the workpiece W while moving along the wavy trajectory TR relative to the workpiece W.

[0023] Next, a method of generating a signal for a weaving operation will be described. First, as a preparation process, an operator operates the input device 38 to input parameters PR of a weaving signal WS for swinging the tool 26 in a weaving operation. The parameters PR include an amplitude value A of the weaving signal WS, a frequency f0 of a fundamental wave FM of the weaving signal WS, and the like. The processor 30 accepts the input of the parameters PR through the input device 38. Thus, in this embodiment, the processor 30 functions as a first input accepting unit 44 (FIG. 2) that accepts an input of the frequency f0 of the fundamental wave FM.

[0024] The processor 30 generates a weaving signal WS in accordance with the parameters PR. Fig. 4 shows a block diagram representing a function for generating the weaving signal WS in the control device 14. The processor 30 functions as an original signal generating unit 46 and generates a weaving signal WS1 in accordance with the received parameters PR.

[0025] An example of the weaving signal WS1 is shown in Fig. 5. The vertical axis of Fig. 5 indicates an amplitude value A (in other words, a distance A from the work path WP) for swinging the tool 26 in the forward and backward directions, and the horizontal axis indicates time t. In the example shown in Fig. 5, the weaving signal WS1 is a triangular wave and has an amplitude value A and a frequency f0 (i.e., a period T0=1 / f0) received from an operator.

[0026] Next, the processor 30 functions as a filter unit 48 and performs a filter process FR1 for removing high frequency components on the weaving signal WS1 generated by the original signal generating unit 46. In this embodiment, the filter unit 48 is configured with a digital filter (FIR filter or IIR filter). For example, the filter unit 48 performs a moving average process on the weaving signal WS1 as a digital signal for each filter time τ, thereby executing the filter process FR1 as a low pass filter process.

[0027] Alternatively, the filter unit 48 may be configured to perform the filtering FR1 as band-pass filtering or notch filtering. The filter unit 48 performs the filtering FR1 on the weaving signal WS1 to generate the weaving signal WS2.

[0028] An example of the weaving signal WS2 is shown in Fig. 6. For comparison, the weaving signal WS1 is shown by a dotted line in Fig. 6. As shown in Fig. 6, as a result of the filter processing FR1, the weaving signal WS1 of a triangular wave changes into a weaving signal WS2 similar to a waveform of a triangular function (i.e., the frequency characteristic FC changes due to the removal of high-frequency components).

[0029] Thus, in this embodiment, the processor 30 functions as the original signal generating unit 46 and the filter unit 48 to generate the weaving signals WS1 and WS2. Therefore, the original signal generating unit 46 and the filter unit 48 configure a weaving signal generating unit 56 that generates the weaving signals WS1 and WS2. Note that, although the weaving signals WS1 and WS2 are shown as analog signals in Fig. 6, in this embodiment, the weaving signals WS1 and WS2 generated by the weaving signal generating unit 56 are digital signals.

[0030] FIG. 7 shows the frequency characteristic FC of the weaving signal WS2 generated by the weaving signal generating unit 56. WS 7 shows a frequency spectrum of the weaving signal WS2. As shown in FIG. 7, the weaving signal WS2 has a fundamental wave FM of frequency f0, and also has a harmonic HM1 of frequency f1, a harmonic HM2 of frequency f2, . . . , a harmonic HM3 of frequency f3, . . . n Harmonics of HM n The frequency components are:

[0031] harmonic HM n Frequency f n may be, for example, an integer multiple (specifically, an odd multiple) of the frequency f0 of the fundamental wave FM. In the example shown in FIG. 7, the fundamental wave FM has an amplitude A0, the harmonic HM1 has an amplitude A1, and the harmonic HM2 has an amplitude A2. Thus, the weaving signal WS2 has a frequency f n (n=0,1,2,3...) frequency components (i.e., the fundamental FM and the harmonics HM n ).

[0032] 2 and 4 again, the processor 30 functions as a main operation command generating unit 50 to generate a main operation command CM1 for causing the robot 12 to move the tool 26 in the direction of the work path WP in the weaving operation described above. Then, the processor 30 functions as a filter unit 52 to perform filter processing FR2 on the main operation command CM1 generated by the main operation command generating unit 50.

[0033] The filter unit 52 is composed of a digital filter similar to the filter unit 48 described above, and performs a filter process FR2 for removing high-frequency components from the main operation command CM1. The filter process FR2 may be the same as or different from the filter process FR1 described above. The filter unit 52 generates the main operation command CM2 by performing the filter process FR2 on the main operation command CM1.

[0034] Next, the processor 30 functions as a weaving motion command generating unit 54, and applies the weaving signal WS2 to the main motion command CM2 to generate a weaving motion command CM3 for causing the robot 12 to perform a weaving motion. Specifically, the weaving motion command generating unit 54 is an adder, and generates a weaving motion command CM3 (=CM2+WS2) by adding the weaving signal WS2 to the main motion command CM2.

[0035] The weaving motion command CM3 generated by the weaving motion command generating unit 54 is output to the servo motor 28 of the robot 12 through the I / O interface 34 and a servo amplifier (not shown). In accordance with this weaving motion command CM3, the robot 12 executes a weaving motion and moves the tool 26 along the trajectory TR shown in FIG.

[0036] The original signal generating unit 46, the main operation command generating unit 50, the filter units 48 and 52, and the weaving operation command generating unit 54 may be functional modules realized by the operation program PG executed by the processor 30. Note that at least one of the main operation command generating unit 50, the filter units 48 and 52, and the weaving operation command generating unit 54 (for example, the filter units 48 and 50) may be implemented in the control device 14 as an analog circuit.

[0037] When the robot 12 is made to perform a weaving operation in accordance with the weaving operation command CM3 generated as described above, the natural frequency of the robot 12 that swings the tool 26 in the forward and backward directions is equal to or exceeds the frequency f0 of the fundamental wave FM of the weaving signal WS2 described in FIG. 7 or the frequency f1 of the harmonic HM.n Frequency f n , so that the robot 12 may resonate.

[0038] FIG. 8 shows a time-varying characteristic A of the amplitude value A in the front-rear direction of the tool 26 (i.e., the y-axis direction of the robot coordinate system C1) when the robot 12 resonates at the frequency f0 of the fundamental wave FM. R1 8, for comparison, the time change characteristics of the weaving signal WS2 generated by the weaving signal generating unit 56 are indicated by a dotted line.

[0039] In addition, Fig. 9 shows the time-varying characteristic A R1 Frequency characteristics of FC R1 As shown in FIG. 9, the frequency characteristic FC R1 is the fundamental FM and harmonic HM contained in the original weaving signal WS2. n However, when the robot 12 resonates at the frequency f0, the amplitude value A0' of the fundamental wave FM becomes greater than the amplitude value A0 of the original weaving signal WS2.

[0040] On the other hand, FIG. 10 shows a time-varying characteristic A of the amplitude value A of the tool 26 in the forward and backward directions when the robot 12 resonates at the frequency f2 of the harmonic HM2. R2 FIG. 11 shows the time-varying characteristic A R2 Frequency characteristics of FC R2 11, when the robot 12 resonates at a frequency f2, the amplitude value A2' of the harmonic HM2 is greater than the amplitude value A2 of the original weaving signal WS2.

[0041] 8 to 11 occurs, the trajectory TR of the tool 26 deviates from the trajectory TR corresponding to the original weaving signal WS2 as shown in Fig. 3, resulting in a decrease in the accuracy of the welding work. Therefore, in this embodiment, when the processor 30 detects such resonance, it corrects the weaving signal WS so as to suppress the resonance.

[0042] A method for correcting the weaving signal WS will be described below. As shown in Fig. 4, in this embodiment, each servo motor 28 is provided with a rotation detector 58 (encoder, Hall element, or the like) that detects the rotation (rotation angle or rotation position) of the servo motor 28.

[0043] Each rotation detector 58 continuously (e.g., periodically) detects the rotation of the servo motor 28 while the robot 12 is performing a weaving operation in accordance with the weaving operation command CM3, and sequentially supplies the rotation to the control device 14 as position feedback FB. c The rotation of the servo motor 28 is periodically detected (for example, every 100 [msec]) and supplied to the control device 14 as position feedback FB.

[0044] The processor 30 continuously acquires the position feedback FB from the rotation detector 58 through the I / O interface 34. Then, the processor 30 functions as a time-series data acquisition unit 60 and acquires time-series data TD of the amplitude value A of the tool 26 in the forward / rearward direction based on the acquired position feedback FB.

[0045] Specifically, each time the processor 30 acquires position feedback FB from the rotation detector 58, it uses the position feedback FB to calculate the coordinates of the tool 26 (specifically, the origin: TCP of the tool coordinate system C2) in the robot coordinate system C1, and obtains the amplitude value A from the coordinates.

[0046] The processor 30 receives the position feedback FB from the rotation detector 58 (i.e., every control period T c ) to obtain the amplitude value A and store it as time series data TD in the memory 32. The time series data TD thus obtained is, for example, the time change characteristic A in FIG. R1 , or time-varying characteristic A in FIG. R2 is a digital signal corresponding to

[0047] Next, the processor 30 functions as a frequency characteristic acquisition unit 62 to obtain the frequency characteristic FC of the time series data TD acquired by the time series data acquisition unit 60. TD Specifically, the processor 30 calculates the first frequency characteristic by dividing the acquired time series data TD by a predetermined period T f Then, the signal is converted to frequency domain data by Fourier transform (e.g., FFT), and the frequency characteristic FC TD Get the.

[0048] The frequency characteristic FC thus obtained TD For example, the frequency characteristic FC shown in FIG. R1 , or the frequency characteristic FC shown in FIG. R2 Note that the period of the Fourier transform is T f For example, the control period T c The period can be arbitrarily set by the operator as an integer multiple of the period T f may be input as the above-mentioned parameter PR.

[0049] Next, the processor 30 functions as a resonance determination unit 64 and determines the frequency characteristic FC acquired by the frequency characteristic acquisition unit 62. TD Based on this, the frequency f of the weaving signal WS2 generated by the weaving signal generating unit 56 is n It is determined whether or not the robot 12 is resonating. A specific example of a method for determining resonance will be described below.

[0050] As a specific example of the resonance determination, the resonance determination unit 64 determines the frequency characteristic FC of the time series data TD acquired by the time series data acquisition unit 60. TD and the frequency characteristic FC of the weaving signal WS2 generated by the weaving signal generating unit 56. WS (second frequency characteristic) to determine whether the robot 12 is resonating.

[0051] Specifically, the processor 30 functions as a frequency characteristic acquisition unit 62 to obtain the above-mentioned frequency characteristic FC TD Also, the frequency characteristic FC of the weaving signal WS2WS For example, the frequency characteristic acquisition unit 62 acquires the weaving signal WS2 generated by the filter unit 48, and converts the weaving signal WS2 into a signal having a predetermined period T f Then, the signal is converted to frequency domain data by Fourier transform (e.g., FFT), and the frequency characteristic FC WS In this way, the frequency characteristic acquisition unit 62 acquires the frequency characteristic FC as shown in FIG. WS Get the.

[0052] As an example, the resonance determination unit 64 determines the frequency characteristic FC of the acquired weaving signal WS2. WS Based on this, the frequency characteristic FC WS The frequency f included in n Amplitude value A at n (the second amplitude value) according to the threshold α n Specifically, the resonance determination unit 64 sets the acquired frequency characteristic FC WS According to the amplitude value A0 (FIG. 7) at the frequency f0 of the fundamental wave FM included in the frequency domain, the threshold value α0 is set as α0=β×A0 (β is a coefficient equal to or greater than 1).

[0053] On the other hand, the resonance determination unit 64 detects the frequency characteristic FC of the acquired time series data TD. TD From the above, the frequency characteristic FC TD The resonance determination unit 64 acquires an amplitude value A0' (first amplitude value) at the frequency f0 of the fundamental wave FM included in the weaving signal WS2. The resonance determination unit 64 then compares the amplitude value A0' with a threshold value α0 to determine whether the amplitude value A0' exceeds the threshold value α0 (i.e., A0'>α0), and determines that the robot 12 is resonating at the frequency f0 of the fundamental wave FM of the weaving signal WS2 if A0'>α0.

[0054] Alternatively, the resonance determination unit 64 may determine the frequency characteristic F. WS The harmonics contained in n Frequency f n Amplitude value A at n (For example, the amplitude values ​​A1, A2, etc. shown in FIG. 7) are used to determine the threshold value α n α n =β×A nThe resonance determination unit 64 may determine the frequency characteristic FC of the acquired time series data TD. TD The harmonics contained in n Frequency f n Amplitude value A at n ' (the first amplitude value, for example, the amplitude value A2' shown in FIG. 11) is equal to or smaller than the threshold value α n Exceeded (A n '>α n ) and A n '>α n If so, the robot 12 detects the harmonic HM n Frequency f n It is determined that resonance occurs at

[0055] The resonance determination unit 64 determines the frequency characteristic F.C. WS The fundamental wave FM and the harmonic HM included in n Frequency f n The amplitude value A of each n For threshold α n Set the frequency characteristic FC TD The fundamental wave FM and the harmonic HM included in n Frequency f n The amplitude value A of each n ' may be compared with any frequency f n The amplitude value A n For threshold α n It may be predetermined by an operator whether to set the value.

[0056] As another example, the resonance determination unit 64 detects the frequency characteristic F.C. WS A given frequency band [f i ,f j ] (for example, the amplitude value A in the frequency band [f0,f2] = f0≦f≦f2) n The integral value of INT(=∫A n ) is calculated, and a threshold value γ is set for the calculated integral value INT as γ=β×INT.

[0057] On the other hand, the resonance determination unit 64 determines the frequency characteristic F.C. TD The frequency band [f i ,f j]Amplitude value A n 'Integral value INT'(=∫A n Then, the resonance determination unit 64 determines whether the acquired integral value INT' exceeds a threshold value γ (INT'>γ), and if INT'>γ, determines that the robot 12 is resonating at any of the frequency components of the weaving signal WS2.

[0058] As yet another example, the resonance determination unit 64 detects the frequency characteristic F.C. WS The amplitude value A n The resonance determination unit 64 calculates the spectral density D (power spectral density, energy spectral density, etc.) of the frequency characteristic FC and sets a threshold value δ for the spectral density D as δ=β×D. TD The amplitude value A n Then, the resonance determination unit 64 determines whether the determined spectral density D' exceeds a threshold value δ (D'>δ), and if D'>δ, determines that the robot 12 is resonating.

[0059] The resonance determination unit 64 determines the frequency characteristic F.C. WS The calculated value (for example, the amplitude value A n , integral value INT or spectral density D) and frequency characteristic FC TD The calculated value (for example, the amplitude value A n Alternatively, the difference Δ between the signal intensity Δ and the signal intensity Δ (signal intensity Δ, integral value INT', or spectral density D') may be calculated.

[0060] Then, the resonance determination unit 64 checks whether the difference Δ is equal to or smaller than a predetermined threshold Δ th Exceeded (Δ>Δ th ) and determine whether Δ > Δ th If so, the robot 12 may be determined to be resonating. As described above, the resonance determination unit 64 determines that the frequency characteristic FC TD and F.C. WS By comparing these with each other, it is determined whether the robot 12 is resonating or not.

[0061] The resonance determination unit 64 determines the above-mentioned threshold value α n , γ, δ or Δth For example, the operator may operate the input device 38 to input the above-mentioned coefficient β or threshold Δ as the threshold setting data THD. th The processor 30 receives the threshold setting data THD, and functions as the resonance determination unit 64 to determine the threshold α n , γ, δ or Δ th That is, in this case, the processor 30 functions as a second input receiving unit 66 (FIG. 2) that receives an input of the threshold setting data THD.

[0062] As another specific example of the resonance determination, the resonance determination unit 64 determines the frequency characteristic FC WS Specifically, the processor 30 functions as a resonance determination unit 64 and can determine whether or not the robot 12 is resonating without acquiring the frequency characteristic FC of the time series data TD acquired by the frequency characteristic acquisition unit 62. TD The fundamental wave FM or harmonic HM included in n Frequency f n Amplitude value A at n ' is a predetermined threshold value α n ' is exceeded.

[0063] Then, the resonance determination unit 64 determines A n '>α n If the threshold value α n ' is the frequency f of the weaving signal WS2 n The frequency characteristic FC of the weaving signal WS2 is determined in advance by an operator based on the above and stored in advance in the memory 32. For example, the operator may determine the frequency characteristic FC of the weaving signal WS2 by using an experimental method or the like. WS By verifying in advance, the threshold α n In this case, the processor 30 functions as the second input receiving unit 66 and receives the threshold value α n ' may be accepted in advance.

[0064] By the above method, the resonance determination unit 64 determines whether or not the robot 12 is resonating. When the resonance determination unit 64 determines that the robot 12 is resonating, the processor 30 functions as a correction unit 68 and corrects the frequency f of the weaving signal WS. n The weaving signal WS is corrected so as to change

[0065] As an example, the correction unit 68 changes the frequency f0 of the fundamental wave FM input as the parameter PR. Specifically, the correction unit 68 changes the frequency f0 set in the parameter PR to a frequency f0' (=f0+φ or f0-φ) by increasing or decreasing the frequency f0 by a predetermined change amount φ. After changing to the frequency f0', the original signal generation unit 46 generates a weaving signal WS1 having the changed frequency f0'.

[0066] As a result, the frequency f0 of the fundamental wave FM of the weaving signal WS2 generated by the filter unit 48 is also changed to the frequency f0'. In this way, the correction unit 68 corrects the weaving signals WS1 and WS2 generated by the weaving signal generation unit 56 by changing the frequency f0 of the fundamental wave FM. In this case, not only the frequency f0 of the fundamental wave FM of the weaving signal WS2 generated by the weaving signal generation unit 56 but also the frequency f0 of the harmonic HM n Frequency f n can also be changed.

[0067] As another example, the resonance determination unit 64 detects the harmonic HM of the weaving signal WS2. n Frequency f n When it is determined that the robot 12 is resonating, the correction unit 68 corrects the harmonic HM n Frequency f n For example, the frequency characteristic acquisition unit 62 may correct the weaving signal WS2 by changing the frequency characteristic FC WS and F.C. TD The resonance determination unit 64 obtains the frequency characteristic F WS and F.C. TDAs a result of the comparison, it is determined that the amplitude value A2' at the frequency f2 of the harmonic HM2 exceeds the threshold value α2, as shown in FIG.

[0068] In this case, the correction unit 68 corrects the frequency characteristic FC of the weaving signal WS2 acquired by the frequency characteristic acquisition unit 62. WS The frequency f2 of the harmonic HM2 included in the weaving signal WS2 is increased or decreased by a predetermined change amount φ by digital signal processing to change it to a frequency f2'. The correction unit 68 then calculates a frequency characteristic FC having the harmonic HM2 of the changed frequency f2' and the fundamental wave FM and harmonics HM1, HM3, HM4, ... of the original weaving signal WS2. WS ' is required.

[0069] The weaving signal generating unit 56 calculates the frequency characteristic FC obtained by the correction unit 68. WS The correction unit 68 performs an inverse Fourier transform (for example, IFFT) on the harmonic HM′ to generate a corrected weaving signal WS2 and outputs the corrected weaving signal WS2 to the weaving operation command generating unit 54. n Frequency f n The weaving signal WS2 can be corrected by changing

[0070] In addition, the resonance determination unit 64 detects the harmonic HM n Frequency f n The amplitude value A n ' is a predetermined threshold α n Even when the resonance of the robot 12 is judged by judging whether the harmonic HM n Frequency f n The weaving signal WS2 can be corrected by changing

[0071] Specifically, it is assumed that the resonance determination unit 64 determines that the amplitude value A2' at the frequency f2 of the harmonic HM2 exceeds a predetermined threshold value α2' as shown in Fig. 11. At this time, the frequency characteristic acquisition unit 62 obtains the frequency characteristic FC WS The correction unit 68 obtains the frequency characteristic F. WSThe frequency f2 of the harmonic HM2 included in the harmonic HM2 is increased or decreased by a change amount φ to change the frequency to f2'.

[0072] Then, as described above, the correction unit 68 calculates a frequency characteristic FC having the harmonic HM2 of the changed frequency f2' and the fundamental wave FM and harmonics HM1, HM3, HM4, . . . of the original weaving signal WS2. WS ' and performs an inverse Fourier transform (e.g., IFFT) to generate a corrected weaving signal WS2. n Frequency f n The weaving signal WS2 can be corrected by changing

[0073] The processor 30 may function as the first input receiving unit 44 and receive the input of the above-mentioned change amount φ along with the frequency f0 of the fundamental wave FM. For example, the operator operates the input device 38 to input the change amount φ in advance. The weaving signal generating unit 56 generates the weaving signal WS2 in accordance with the received amplitude value A and frequency f0, and the correcting unit 68 corrects the frequency f of the weaving signal WS in accordance with the received change amount φ when the resonance determining unit 64 determines that the robot 12 is resonating. n In this manner, the correction unit 68 changes the frequency f n The weaving signal WS is corrected so as to change

[0074] Next, a control flow of the welding operation executed by the processor 30 will be described with reference to Fig. 12. The flow shown in Fig. 12 is started when the processor 30 receives an operation start command from an operator, a higher-level controller, or the operation program PG. In step S1, the processor 30 starts a weaving operation.

[0075] Specifically, as described above, the processor 30 functions as the weaving signal generator 56 (the original signal generator 46, the filter unit 48), the main motion command generator 50, the filter unit 52, and the weaving motion command generator 54 to generate a weaving motion command CM3 and operate the robot 12 according to the weaving motion command CM3. As a result, the robot 12 starts a weaving motion.

[0076] In step S2, the processor 30 operates the tool 26 to cause the tool 26 to perform a welding operation. In this way, the processor 30 performs a welding operation on the workpiece W using the tool 26 while performing the weaving operation started in step S1. In this way, the processor 30 functions as a tool control unit 70 (FIG. 2) that operates the tool 26 to cause the tool 26 to perform a welding operation. In step S3, the processor 30 functions as the time-series data acquisition unit 60 and starts the operation of acquiring the time-series data TD as described above.

[0077] In step S4, the processor 30 functions as the frequency characteristic acquisition unit 62 to acquire the frequency characteristic FC of the time series data TD as described above. TD For example, the frequency characteristic acquisition unit 62 starts an operation of acquiring the frequency characteristic FC of the weaving signal WS2 generated by the weaving signal generation unit 56. WS and the frequency characteristic FC of the time series data TD acquired by the time series data acquisition unit 60 when the robot 12 is operated in accordance with the weaving signal WS2. TD and continuously (for example, control period T c ) to obtain.

[0078] In step S5, the processor 30 functions as the resonance determination unit 64 and, as described above, determines the frequency characteristic F C obtained in step S4. TD For example, the processor 30 determines whether the robot 12 is resonating based on the frequency characteristic FC acquired in the most recent step S4. WS and F.C. TDBy comparing these, it is determined whether or not the robot 12 is resonating. If the processor 30 determines that the robot 12 is resonating (i.e., YES), the process proceeds to step S6, whereas if the processor 30 determines that the robot 12 is resonating (NO), the process proceeds to step S7.

[0079] In step S6, the processor 30 functions as the correction unit 68 to correct the frequency f of the weaving signal WS as described above. n For example, the processor 30 corrects the weaving signal WS so as to change the frequency f0 of the fundamental wave FM of the weaving signal WS1 (or the frequency f1 of the harmonic HM of the weaving signal WS2). n Frequency f n ) by the change amount φ to change it to frequency f0′. After that, the processor 30 functions as the weaving signal generator 56 and generates the weaving signals WS1 and WS2 based on the changed frequency f0′.

[0080] After step S6, the processor 30 returns to step S5 and repeatedly executes the loop of steps S5 and S6 until it is determined to be NO in step S5. Here, the processor 30 increases the frequency f of the weaving signal WS every time it executes step S6. n within a predetermined tolerance range [f th1 ,f th2 ], the change may be made by an amount of change φ.

[0081] Tolerance [f th1 ,f th2 ](=f th1 ≦f n '≦f th2 ) can be arbitrarily determined by the operator as a range of [3.5, 4.5] (i.e., 4 [Hz] ±0.5 [Hz]) when the frequency f0 of the fundamental wave FM is f0 = 4 [Hz]. In this case, the processor 30 functions as the first input receiving unit 44 and receives the allowable range [f th1 ,f th2 ], a threshold value f th1 and f th2 may be accepted.

[0082] By executing step S6, the changed frequency f n ' is within the allowable range [f th1 ,f th2 ] outside the range (i.e., f n ' <f th1 , or f th2 <f n '), and a warning signal indicating that may be generated as image data (or audio data) and displayed on the display device 40 (or output through a speaker). n ' is within the allowable range [f th1 ,f th2 When the welding operation is outside the predetermined range, a warning signal may be output and the welding operation may be continued, or a warning signal may be output and the welding operation may be stopped.

[0083] In step S7, the processor 30 determines whether or not the welding operation has been completed based on the work program PG and the feedback FB from the rotation detector 58. When the processor 30 determines that the welding operation has been completed (i.e., YES), the processor 30 functions as the tool control unit 70 to stop the operation of the tool 26 and ends the flow shown in Fig. 12. On the other hand, when the processor 30 determines that the operation has been completed, the processor 30 returns to step S5.

[0084] 12, the processor 30 may start step S1 after starting step S2. Alternatively, the operator may not input the change amount φ in advance, and the processor 30 may change the frequency f n Alternatively, the amount of change φ by which the value of the vector φ is changed may be determined randomly.

[0085] As described above, in this embodiment, the processor 30 functions as the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determining unit 64, the second input receiving unit 66, the correction unit 68, and the tool control unit 70, and generates signals for the weaving operation.

[0086] Therefore, the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determining unit 64, the second input receiving unit 66, the correcting unit 68, and the tool control unit 70 constitute a device 80 (FIG. 2) that generates a signal for a weaving operation. That is, in this embodiment, the device 80 is implemented in the control device 14 as a functional module realized by the processor 30 executing the work program PG.

[0087] As described above, the device 80 includes the weaving signal generating unit 56 that generates the weaving signals WS1 and WS2, the time series data acquiring unit 60 that acquires the time series data TD when the robot 12 is caused to perform a weaving operation in accordance with the weaving signal WS2, and the frequency characteristic FC of the time series data TD. TD and a frequency characteristic acquisition unit 62 for acquiring a (first frequency characteristic).

[0088] The device 80 also has a frequency characteristic FC TD a resonance determination unit 64 for determining whether the robot 12 is resonating based on the frequency f n and a correction unit 68 for correcting the weaving signal WS1 or WS2 so as to change the weaving signal WS1 or WS2.

[0089] According to this configuration, it is possible to suppress resonance that occurs when the tool 26 is swung back and forth during the weaving operation, and therefore it is possible to perform the weaving operation more stably. As a result, it is possible to increase the accuracy of the welding work, thereby improving the welding quality.

[0090] In the device 80, the correction unit 68 corrects the weaving signals WS1 and WS2 by, for example, changing the frequency f0 of the fundamental wave FM of the weaving signals WS1 and WS2. According to this configuration, as described above, not only the frequency f0 of the fundamental wave FM of the weaving signal WS2 but also the frequency f0 of the harmonic HM n Frequency f nTherefore, the fundamental FM or harmonic HM of the weaving signal WS2 can be changed. n Any frequency f n Even if the robot 12 resonates, the resonance can be easily and reliably suppressed.

[0091] Moreover, in the device 80, the first input receiving unit 44 receives an input of the frequency f0 of the fundamental wave FM, the weaving signal generating unit 56 generates weaving signals WS1, WS2 having the received frequency f0, and the correcting unit 68 corrects the weaving signals WS1, WS2 generated by the weaving signal generating unit 56 by changing the received frequency f0.

[0092] According to this configuration, the operator can arbitrarily set the frequency f0, making it possible to design a greater variety of weaving movements, and by changing the frequency f0 set by the operator, it is possible to reliably suppress resonance of the robot 12 caused by the frequency f0.

[0093] In the device 80, the first input receiving unit 44 receives the fundamental wave FM or the harmonic HM. n Frequency f n The correction unit 68 further receives an input of a change amount φ for changing the frequency f n According to this configuration, the operator changes the frequency f n The change amount φ can be determined arbitrarily so as to reliably suppress resonance while minimizing the effect of the change on the welding quality.

[0094] In the device 80, for example, the frequency characteristic acquisition unit 62 acquires the frequency characteristic FC of the weaving signal WS2 generated by the weaving signal generation unit 56. WS The resonance determination unit 64 further acquires the frequency characteristic FC acquired by the frequency characteristic acquisition unit 62. TD and F.C. WSBy comparing these, it is determined whether or not the robot 12 is resonating. According to this configuration, it is possible to detect with higher accuracy whether or not the robot 12 is resonating.

[0095] In addition, the frequency characteristic FC TD and F.C. WS In comparing the frequency characteristics F C TD A given frequency f n The amplitude value A n '(first amplitude value) is the frequency characteristic FC WS The predetermined frequency f n The amplitude value A n (2nd amplitude value) n If it exceeds this, it is determined that the robot 12 is resonating. According to this configuration, it is possible to detect with high accuracy whether or not the robot 12 is resonating, using a relatively simple algorithm.

[0096] In addition, the second input receiving unit 66 sets a threshold value α n The resonance determination unit 64 receives an input of threshold setting data THD (coefficient β) for setting the amplitude value A n For threshold α n According to this configuration, the operator sets the threshold value α n can be set arbitrarily.

[0097] On the other hand, as another example, the resonance determination unit 64 may determine the frequency characteristic FC of the time series data TD acquired by the frequency characteristic acquisition unit 62. TD A given frequency f n The amplitude value A n ' is the frequency f of the weaving signal WS2. n A predetermined threshold value α based on n If it exceeds the threshold value, it is determined that the robot 12 is resonating.

[0098] According to this configuration, it is possible to detect whether or not the robot 12 is resonating with a simpler algorithm. In this case, the second input receiving unit 66 sets the threshold value α nAccording to this configuration, the operator can input the threshold value α n ' can be set arbitrarily.

[0099] Also, according to the device 80, in the flow shown in FIG. 12, the weaving signal generating unit 56 generates the weaving signals WS1, WS2 when the tool control unit 70 is performing the welding operation (step S2), and the time series data acquiring unit 60 acquires the time series data TD when the weaving operation is performed in accordance with the weaving signals WS1, WS2 generated when the welding operation is performed (step S3).

[0100] The resonance determination unit 64 detects the frequency characteristic FC acquired by the frequency characteristic acquisition unit 62 during the welding operation. TD Based on the result of the above (step S4), it is determined whether or not the robot 12 is resonating (step S5), and the correction unit 68 corrects the weaving signals WS1 and WS2 when it is determined that the robot 12 is resonating during the execution of the welding operation (step S6). According to this configuration, it is possible to monitor in real time whether or not the robot 12 is resonating while the welding operation is actually being performed, and to correct the weaving signal WS in real time when the resonance occurs.

[0101] Incidentally, before actually performing welding work (for example, when teaching the robot 12 a weaving operation), it is also possible to verify the occurrence of resonance in the robot 12 and the operation of correcting the weaving signal WS. Hereinafter, a robot system 100 that performs such a function will be described with reference to Figs. 13 and 14.

[0102] The robot system 100 differs from the above-described robot system 10 in that the robot system 100 further includes a teaching device 102. The teaching device 102 teaches the robot 12 a weaving operation. Specifically, the teaching device 102 is, for example, a portable computer such as a teaching pendant or a tablet terminal device, and includes a processor 104, a memory 106, and an I / O interface 108. The teaching device 102 also includes an input device 110 and a display device 112.

[0103] The configurations and functions of the processor 104, memory 106, I / O interface 108, input device 110, and display device 112 are similar to those of the above-mentioned processor 30, memory 32, I / O interface 34, input device 38, and display device 40, so duplicated explanations will be omitted. The processor 104 is communicatively connected to the memory 106 and I / O interface 108 via a bus 114, and performs calculations for executing a teaching function, which will be described later, while communicating with these components.

[0104] In this embodiment, the functions of the device 80 (the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determination unit 64, the second input receiving unit 66, the correction unit 68, and the tool control unit 70) are implemented in the teaching device 102, and the processor 104 of the teaching device 102 functions as the device 80.

[0105] The processor 104 is configured to send commands to the servo motors 28 of the robot 12 via the control device 14 in response to input data to the input device 110, and to jog the robot 12 in accordance with the commands. The operator jogs the robot 12 by operating the input device 110, and teaches the robot 12 a weaving motion.

[0106] Specifically, the operator operates the input device 110 to teach the robot 12 a plurality of teaching points TP for moving the tool 26 (TCP) in the direction of the work path WP. Meanwhile, the operator operates the input device 110 to input parameters PR (amplitude value A, frequency f0, etc.) of the weaving signal WS.

[0107] The processor 104 acquires the coordinates of the teaching point TP in the robot coordinate system C1, and functions as the first input receiving unit 44 to receive the input of the parameters PR. The processor 104 then generates a work program PG1 based on the acquired teaching point TP and parameters PR. In this work program PG1, the coordinates of the teaching point TP and the parameters PR are defined as command codes. In this manner, the teaching device 102 teaches the robot 12 a weaving operation, and creates a work program PG1 for welding work.

[0108] The operator makes the robot 12 try to perform a weaving operation according to the created work program PG1, and verifies whether the robot 12 resonates or not. Fig. 15 shows a block diagram showing the function of generating the weaving signal WS in the teaching device 102. The operator operates the input device 110 to give a work start command to the processor 104.

[0109] In response to the work start command, the processor 104 functions as the weaving signal generating unit 56 to generate the weaving signal WS, similarly to the above-described embodiment. Specifically, the processor 104 functions as the original signal generating unit 46 to generate the weaving signal WS1 according to the parameters PR defined in the work program PG1, and functions as the filter unit 48 to perform the filter processing FR1 on the weaving signal WS1 generated by the original signal generating unit 46, thereby generating the weaving signal WS2.

[0110] Furthermore, similarly to the above-described embodiment, the processor 104 functions as a main operation command generating unit 50 to generate a main operation command CM1 in accordance with the teaching point TP defined in the work program PG1, and functions as a filter unit 52 to perform filter processing FR2 on the main operation command CM1 generated by the main operation command generating unit 50 to generate a main operation command CM2.

[0111] Then, the processor 104 functions as the weaving motion command generating unit 54, generates a weaving motion command CM3 by applying the weaving signal WS2 to the main motion command CM2, and transmits it to the servo motor 28 of the robot 12 via the I / O interface 108 and the control device 14. In this way, the processor 104 controls the robot 12 in accordance with the weaving motion command CM3 via the control device 14, causing it to perform a weaving motion.

[0112] Here, in this embodiment, while the robot 12 is performing the weaving operation, the processor 104 functions as the tool control unit 70 and stops the operation of the tool 26. Therefore, the robot 12 performs the weaving operation by moving the tool 26 along the trajectory TR as shown in FIG. 3 while the operation of the tool 26 is stopped (i.e., without performing welding work).

[0113] While the robot 12 is executing a weaving operation, the rotation detector 58 continuously detects the rotation of the servo motor 28 and supplies it as position feedback FB to the teaching device 102 via the control device 14, similar to the above-described embodiment. Then, the processor 104 functions as a time-series data acquisition unit 60, a frequency characteristic acquisition unit 62, and a resonance determination unit 64, similar to the above-described embodiment, to determine whether or not the robot 12 is resonating.

[0114] In this way, the operator can verify whether or not the robot 12 resonates when the robot 12 executes a weaving operation according to the task program PG1. If the occurrence of resonance is detected, the operator can determine the frequency f at which the robot 12 resonates.n can be identified.

[0115] When resonance of the robot 12 is detected, the operator operates the input device 110 to input the above-mentioned change amount φ. The processor 104 functions as the first input receiving unit 44 to receive the input of the change amount φ, and functions as the correction unit 68 to correct the frequency f n (For example, the frequency f0 of the fundamental wave FM is changed, thereby correcting the weaving signals WS1 and WS2.

[0116] Then, the weaving signal generator 56 generates a corrected weaving signal WS2, and the resonance determiner 64 determines again whether or not resonance occurs in the robot 12 when the robot 12 is caused to perform a weaving operation in accordance with the corrected weaving signal WS2. By repeating the above-described trial process, the operator can obtain an optimal change amount φ that can suppress resonance of the robot 12.

[0117] Thereafter, the processor 104 adjusts the frequency f, which is defined as a parameter PR in the work program PG1, in response to input data from the operator. n by the change amount φ. For example, the processor 104 changes the frequency f0 of the fundamental wave FM defined in the work program PG1 by the change amount φ, and generates a work program PG2 in which the changed frequency f0' is defined as a command.

[0118] The processor 104 supplies the generated operation program PG2 to the control device 14. The processor 30 of the control device 14 operates the robot 12 in accordance with the operation program PG2 to perform an actual welding operation on the workpiece W. Here, according to the operation program PG2, the frequency f n is the frequency f n Since the program PG2 has been changed to ', even if the robot 12 is caused to perform a weaving operation in accordance with the work program PG2, the robot 12 will not resonate. Therefore, the work quality can be improved.

[0119] As described above, in this embodiment, the weaving signal generating unit 56 generates the weaving signals WS1, WS2 when the tool control unit 70 stops the tool 26, and the time series data acquiring unit 60 acquires the time series data TD when the weaving operation is performed in accordance with the weaving signal WS2 generated when the tool 26 is stopped.

[0120] The resonance determination unit 64 then calculates the frequency characteristic FC obtained when the tool 26 is stopped. TD Whether or not the robot 12 is resonating is judged based on the result of the above. In this manner, by verifying the resonance of the robot 12 at the stage of teaching the weaving operation to the robot 12 without operating the tool 26, the occurrence of the resonance during the actual welding work can be reliably suppressed, and the work quality can be improved.

[0121] The resonance determination unit 64 determines the frequency characteristic FC acquired by the frequency characteristic acquisition unit 62. TD The frequency characteristic FC may be applied to a learning model LM constructed by machine learning to determine whether the robot 12 is resonating. TD 4 shows the correlation between the resonance generated in the robot 12 and the vibration generated in the robot 12.

[0122] The learning model LM is, for example, a frequency characteristic FC obtained when the robot 12 executes a weaving operation according to a weaving signal WS having various parameters PR. TD The learning model LM can be constructed by repeatedly providing a learning data set of the resonance data and the judgment data indicating the presence or absence of resonance at that time to the machine learning device (for example, supervised learning). The learning model LM is stored in advance in the memory 32 or 106.

[0123] The resonance determination unit 64 detects the frequency characteristic FC acquired by the frequency characteristic acquisition unit 62 during the weaving operation. TD are input to the learning model LM in sequence. The learning model LM is TDIn this way, the resonance determination unit 64 determines whether or not resonance occurs based on the frequency characteristic F C TD Using the learning model LM, it is possible to determine with high accuracy whether or not the robot 12 is resonating.

[0124] The learning model LM has a frequency characteristic FC TD When the frequency characteristic FC is input, the presence or absence of resonance is judged, and when it is judged that resonance is occurring, an optimal change amount φ is output. TD , the resonance occurring in the robot 12, and the correlation of the change amount φ are shown, and various frequency characteristics FC TD The learning data set can be constructed by repeatedly providing the machine learning device with a learning data set including the judgment data indicating the presence or absence of resonance at that time and the change amount φ associated with the judgment data indicating the occurrence of resonance.

[0125] In this case, the learning model LM is a frequency characteristic FC acquired by the resonance determination unit 64 during the weaving operation. TD When the frequency characteristic FC is input, the presence or absence of resonance is judged, and if it is judged that resonance is occurring, the input frequency characteristic FC TD Then, the correction unit 68 outputs the change amount φ having a high correlation with the frequency f n According to this configuration, when resonance of the robot 12 occurs, the weaving signal WS can be corrected by the optimal change amount φ obtained by machine learning.

[0126] In addition, the processor 30 or 104 may swing the tool 26 in the up-down direction in synchronization with swinging the tool 26 in the forward and backward directions during the weaving operation. Such a weaving operation will be described with reference to Fig. 3 and Fig. 16. In the example shown in Fig. 16, the processor 30 swings the tool 26 in the forward and backward directions, and also swings it in the up-down direction so as to displace it upward by a distance δ at the position of the movement path WP.

[0127] The weaving signal WS1 generated by the weaving signal generating unit 56 to cause the robot 12 to perform a swinging motion in the up and down direction as shown in FIG. z and WS2 z The weaving signal WS1 shown in FIG. 17 indicates the amplitude A of the vertical swing of the tool 26, and the horizontal axis indicates time t. z The parameter PR is the amplitude A and the frequency f 0_z (=1 / T 0_z ).

[0128] As in the above embodiment, the original signal generator 46 generates a weaving signal WS1 z and the filter unit 48 generates the weaving signal WS1 z By applying the filter process FR1 to the weaving signal WS2 z This weaving signal WS2 z When the robot 12 is caused to swing the tool 26 in the vertical direction in accordance with the weaving signal WS2 z Frequency f n_z The robot 12 may resonate.

[0129] The processor 30 or 104 functions as a time series data acquisition unit 60, a frequency characteristic acquisition unit 62, and a resonance determination unit 64, and determines whether or not resonance occurs in the robot 12 in the same manner as in the above-described embodiment. When it is determined that resonance occurs, the processor 30 or 104 determines whether or not resonance occurs in the robot 12 in the same manner as in the above-described embodiment. n_z Change the weaving signal WS1 z Or WS2 z can be corrected.

[0130] In the above embodiment, the processor 30 or 104 determines whether or not the robot 12 resonates and corrects the weaving signal WS when the actual robot 12 executes a weaving operation (so-called online). However, the present invention is not limited to this, and the processor 30 or 104 may function as the device 80 to execute a simulation of the weaving operation of the robot 12 in a virtual space, and in the simulation, determine whether or not the robot 12 resonates and correct the weaving signal WS (so-called offline).

[0131] In the above embodiment, the function of the device 80 is implemented in either the control device 14 or the teaching device 102. However, the present invention is not limited to this, and the function of the device 80 may be implemented in both the control device 14 and the teaching device 102. That is, in this case, the device 80A (the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determining unit 64, the second input receiving unit 66, the correction unit 68, and the tool control unit 70) is implemented in the control device 14, and the device 80B (the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determining unit 64, the second input receiving unit 66, the correction unit 68, and the tool control unit 70) is implemented in the teaching device 102.

[0132] Alternatively, some of the functions of the device 80 (e.g., the correction unit 68 and the tool control unit 70) may be implemented in the control device 14, while other parts of the functions of the device 80 (e.g., the first input receiving unit 44, the weaving signal generating unit 56, the time series data acquiring unit 60, the frequency characteristic acquiring unit 62, the resonance determination unit 64, and the second input receiving unit 66) may be implemented in the teaching device 102.

[0133] The functions of the device 80 may be implemented in any other computer (such as a desktop PC) communicatively connected to the control device 14 or the teaching device 102. In this case, the processor of the other computer may function as the device 80 and control the robot 12 via the control device 14 to perform a weaving operation.

[0134] Also, in the above embodiment, the case has been described in which the tool 26 is a welding torch and performs a welding operation on the workpiece W. However, the present invention is not limited to this, and the tool 26 may be configured to, for example, melt a brazing material fed from a brazing material feeder (not shown) to perform a brazing operation on the workpiece W, or may be configured to perform any operation on the workpiece W.

[0135] Furthermore, the robot 12 is not limited to a vertical articulated robot, but may be any type of robot capable of moving the tool 26, such as a horizontal articulated robot or a parallel link robot. The present disclosure has been described above through the embodiments, but the above-mentioned embodiments do not limit the invention according to the claims. The present disclosure has been described above through the embodiments, but the above-mentioned embodiments do not limit the invention according to the claims. [Explanation of symbols]

[0136] 10,100 Robot Systems 12. Robot 14 Control device 26 Tools 30,104 processors 44,66 Input reception section 56 Weaving signal generator 60 Time series data acquisition section 62 Frequency characteristic acquisition section 64 Resonance determination section 68 Correction section 70 Tool control section 80 equipment 102 Teaching device

Claims

1. 1. An apparatus for generating a signal for a weaving operation in which a robot moves a tool along a predetermined work path and swings the tool in a direction intersecting the work path, the apparatus comprising: a weaving signal generating unit that generates a weaving signal for causing the robot to swing the tool; a time series data acquiring unit that acquires time series data of an amplitude value of the tool in the intersecting direction when the robot is caused to perform the weaving operation in accordance with the weaving signal generated by the weaving signal generating unit; a frequency characteristic acquisition unit that acquires a first frequency characteristic of the time series data acquired by the time series data acquisition unit; a resonance determination unit that determines whether or not the robot resonates at a frequency of the weaving signal generated by the weaving signal generation unit, based on the first frequency characteristic acquired by the frequency characteristic acquisition unit; a correction unit that corrects the weaving signal to change the frequency when the resonance determination unit determines that the robot is resonating.

2. The apparatus of claim 1 , wherein the correction unit corrects the weaving signal by changing the frequency of a fundamental wave of the weaving signal.

3. a first input receiving unit that receives an input of the frequency of the fundamental wave, the weaving signal generation unit generates the weaving signal having the frequency accepted by the first input acceptance unit, The device according to claim 2 , wherein the correction unit corrects the weaving signal generated by the weaving signal generation unit by changing the frequency accepted by the first input acceptance unit.

4. The first input receiving unit further receives an input of an amount of change for changing the frequency, The device according to claim 3 , wherein the correction unit changes the frequency in accordance with the amount of change accepted by the first input acceptance unit.

5. The frequency characteristic acquisition unit further acquires a second frequency characteristic of the weaving signal generated by the weaving signal generation unit, The device according to any one of claims 1 to 4, wherein the resonance determination unit determines whether or not the robot is resonating by comparing the first frequency characteristic and the second frequency characteristic acquired by the frequency characteristic acquisition unit.

6. 6. The device according to claim 5, wherein the resonance determination unit determines that the robot is resonating when a first amplitude value of a predetermined frequency included in the first frequency characteristic exceeds a threshold value determined according to a second amplitude value of the predetermined frequency included in the second frequency characteristic.

7. a second input receiving unit that receives an input of threshold setting data for setting the threshold; The apparatus according to claim 6 , wherein the resonance determination unit sets the threshold for the second amplitude value in accordance with the threshold setting data received by the second input receiving unit.

8. The device according to any one of claims 1 to 4, wherein the resonance determination unit determines that the robot is resonating when a first amplitude value of a predetermined frequency included in the first frequency characteristic acquired by the frequency characteristic acquisition unit exceeds a predetermined threshold value based on the frequency of the weaving signal.

9. The apparatus according to claim 8 , further comprising a second input receiving unit that receives an input of the threshold value.

10. a tool control unit that operates the tool to perform a predetermined task; the weaving signal generating unit generates the weaving signal when the tool control unit stops the tool, the time series data acquisition unit acquires the time series data when the weaving operation is performed in accordance with the weaving signal generated by the weaving signal generation unit when the tool is stopped, The apparatus according to any one of claims 1 to 9, wherein the resonance determination unit determines whether or not the robot is resonating based on the first frequency characteristic acquired by the frequency characteristic acquisition unit when the tool is stopped.

11. a tool control unit that operates the tool to perform a predetermined task; the weaving signal generating unit generates the weaving signal when the tool control unit is performing the predetermined task, the time series data acquisition unit acquires the time series data when the weaving operation is performed in accordance with the weaving signal generated by the weaving signal generation unit during the execution of the predetermined work, the resonance determination unit determines whether or not the robot is resonating based on the first frequency characteristic acquired by the frequency characteristic acquisition unit during execution of the predetermined task; The device according to any one of claims 1 to 9, wherein the correction unit corrects the weaving signal when the resonance determination unit determines that the robot is resonating during execution of the specified task.

12. A control device comprising the device according to any one of claims 1 to 11, and configured to control the robot to perform the weaving operation.

13. 1. A method for generating signals for a weaving operation by a robot to move a tool along a predetermined work path and to swing the tool in a direction intersecting the work path, comprising: The processor: generating a weaving signal to cause the robot to wobble the tool; acquiring time series data of amplitude values ​​of the tool in the intersecting direction when the robot is caused to perform the weaving operation in accordance with the generated weaving signal; Acquire a first frequency characteristic of the acquired time series data; determining whether or not the robot resonates at a frequency of the generated weaving signal based on the acquired first frequency characteristic; and amending the weaving signal to change the frequency when the robot is determined to be in resonance.

Citation Information

Patent Citations

  • Method and device for correcting weaving width of industrial robot

    JP1991207576A

  • Robot weaving controller

    JP1994222817A

  • Weaving control device for robot

    JP1998235582A

Cited By

  • Numerical control device and numerical control method

    JP7814645B1