JPH06250723A - Vibration reducing device for robot - Google Patents

Abstract

PURPOSE:To reduce the vibration of the robot and perform real-time processing by integrating the acceleration shown by the acceleration pattern refound by processing the power spectrum distribution, obtained by removing a part corresponding to a frequency at which the robot is caused to vibrate from a power spectrum distribution, by inverse Fourier transformation. CONSTITUTION:When the acceleration/deceleration time tP3 corresponding to a teaching speed omegaT is selected from a table 2a at the time of PTP control, scale conversion parts 3a and 3b read a normalization speed pattern after the resonance frequency component of a machine system is removed out of a normalization speed pattern memory 4. Further, the scale arithmetic parts 3a and 3b converts the taught speed into a command speed by using the read normalized speed pattern. The scale arithmetic part 3a converts the speed into a command speed omega(i) of each axis (i) of the robot. Those command speeds omega(i) are integrated by an integrator 5 to find command positions theta(i) of the respective axes (i), which are outputted to a servo arithmetic part 8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an industrial robot, and more particularly, to an apparatus capable of reducing vibration generated in an arm when the robot starts or stops its operation.

[0002]

2. Description of the Related Art In the field of industrial machines including industrial robots, various control techniques for reducing vibration during machine operation have been developed and applied to machines in which they vibrate. Among these vibration reduction controls, as one of the methods that can be realized relatively easily and have a great effect,
There is a vibration reduction method by so-called acceleration / deceleration processing.

Among the acceleration / deceleration processing vibration reduction methods,
As one of the methods aimed at suppressing the resonance of the mechanical system, there is a method of removing a frequency component near the resonance frequency of the mechanical system from the command value. As the conventional technique of this method,
As disclosed in Japanese Patent Laid-Open No. 62-125411, for the purpose of moving and stopping the actuator of the magnetic disk at high speed and smoothly to a position separated by an arbitrary access distance, as disclosed in the publication. , The drive component near the resonance frequency is removed by changing the ratio of the upper base and the lower base in the trapezoidal wave drive current as a command value given to the actuator.

[0004]

As described above, the conventional technique uses the method of removing the frequency component in the vicinity of the resonance frequency of the mechanical system with respect to the current command value. Can be achieved,
Since a position or velocity command is usually used in an industrial robot, this conventional technique cannot be applied as it is. Moreover, in the case of a robot, the resonance frequency changes depending on its movement position, that is, the posture that the robot can take. Therefore, when processing is performed online, the resonance frequency is calculated and then the resonance frequency near the calculated resonance frequency is calculated. Since the arithmetic processing for changing the ratio of the upper base and the lower base of the trapezoidal wave is performed in order to remove the driving component, the amount of calculation to be performed online increases, and it is impossible to perform the processing in real time. Will also be invited.

The present invention has been made in view of the above circumstances, and provides a vibration reducing apparatus that can be applied to reduce the vibration of a robot and can perform processing in real time without increasing the amount of calculation. That is the purpose.

[0006]

Therefore, according to the present invention,
Teach the movement speed of the tool tip of the robot or the driving speed of each axis of the robot, calculate the acceleration time or deceleration time until it reaches the taught speed, and teach the speed within the calculated acceleration time or deceleration time. The speed pattern showing the time change of the speed until reaching is reached, and the robot each axis is driven and controlled until the moving speed of the tool tip or the driving speed of each robot axis reaches the teaching speed according to the obtained speed pattern. In the robot configured as described above, an acceleration pattern showing the time change of the acceleration until reaching the teaching speed is prepared, the acceleration shown in this acceleration pattern is Fourier transformed to obtain the power spectrum distribution, and the robot is calculated from this power spectrum distribution. Remove the part corresponding to the frequency that excites the vibration,
The remaining power spectrum distribution thus removed is subjected to inverse Fourier transform to recalculate the acceleration pattern, and the speed pattern is calculated by integrating the acceleration shown in the recalculated acceleration pattern.

[0007]

That is, in the present invention, attention is paid to the fact that the acceleration when the tool tip of the robot or each axis of the robot moves substantially corresponds to the driving force acting on the arm of the robot. The resonance frequency that excites the vibration is known. Therefore, an acceleration pattern showing the time change of the acceleration until reaching the taught velocity is prepared, and the power spectrum distribution is obtained by Fourier transforming the acceleration shown in this acceleration pattern, and the vibration is excited in the robot from this power spectrum distribution. The portion corresponding to the frequency to be removed is removed, the remaining power spectrum distribution thus removed is subjected to inverse Fourier transform to recalculate the acceleration pattern, and the speed pattern is calculated by integrating the acceleration shown in the recalculated acceleration pattern. Is calculated. Then, according to the calculated speed pattern, each robot axis is drive-controlled until the moving speed of the tool tip or the driving speed of each robot axis reaches the teaching speed.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A vibration reducing apparatus for a robot according to the present invention will be described below with reference to the drawings.

FIG. 1 is a control block diagram of an embodiment apparatus for controlling the vibration of a robot.

As shown in the figure, this device can support PTP control or CP control as the control of the trajectory of the robot, and indicates the teaching speed indicating the angular velocity of each axis of the robot or the moving velocity of the tool tip of the robot. The teaching speed is given to the acceleration / deceleration time selection processing unit 1a for PTP control or the acceleration / deceleration time selection processing unit 1b for CP control. Then, the selection processing unit 1a or 1b refers to the contents of the acceleration / deceleration time selection table 2a for PTP control or the acceleration / deceleration time selection table 2b for CP control to determine the optimum acceleration until reaching the given teaching speed. Deceleration time is selected. Here, in each of the tables 2a and 2b, as will be described later, the acceleration / deceleration time corresponding to the teaching speed is stored in advance. For example, in the case of PTP control, the acceleration / deceleration time tp3 corresponding to the teaching speed ωT is selected from the table 2a.

When the acceleration / deceleration time is selected in this way, the scale conversion calculation units 3a and 3b use the stored contents of the normalized speed pattern memory 4 which is shared by the PTP control and the CP control, as will be described later. The normalized velocity pattern from which the resonance frequency component of the system has been removed is read. Here, it is assumed that a normalized speed pattern corresponding to the acceleration / deceleration time is stored in advance in the memory 4, as described later.
Further, in the scale calculators 3a and 3b, the teaching speed is converted into a command speed using the read normalized speed pattern. For example, in the case of PTP control, the scale calculator 3a converts the command speed ω (i) for each axis i of the robot and outputs the command speed ω (i). And this converted command speed ω
(I) is integrated in the integrator 5 and the command position θ of each axis i
(I) is obtained and output to the servo calculation unit 8.

On the other hand, in the case of CP control, the scale calculator 3b calculates the moving speed of the tool tip of the robot, and the target trajectory calculator 6 calculates the target trajectory of the tool tip. Then, on the basis of this target trajectory, the command positions of the respective axes of the robot are obtained in the inverse transformation calculation unit 7,
It is output to these servo calculation units 8.

The servo calculation unit 8 drives and controls each axis of the robot based on the input axis command position according to the read acceleration pattern until the teaching speed is reached.

Now, in order to understand the embodiment, the acceleration / deceleration processing which the present applicant has already adopted will be described with reference to the control block diagram of FIG.

Now, it is assumed that PTP control is performed and acceleration processing is performed.

When the teaching speed is given to the acceleration / deceleration time calculation unit 11, the calculation unit 11 calculates the acceleration time until the teaching speed is reached, and outputs the calculated acceleration time to the scale conversion unit 12. In the conversion unit 12, the normalized speed pattern is read from the normalized speed pattern memory 13, and the command speed is converted and obtained based on the read normalized speed pattern and the calculated acceleration time. It is output to the unit 14.

Here, the normalized speed pattern stored in the normalized speed pattern memory 13 is, as shown in FIG. 3C, up to the acceleration time tmax when each axis of the robot is accelerated at the maximum speed. It is a change in speed with time, and only speed is normalized.

Therefore, the time axis of FIG. 3 (c) is compressed proportionally to the acceleration time calculated by the acceleration / deceleration time calculation unit 11, and the normalized speed data is multiplied by the teaching speed. The commanded speed will be obtained. In the case of the deceleration process, the command speed is obtained by performing the same process as the above-described acceleration process using a pattern in which FIG. 3C is symmetrical with respect to the time axis.

The command speed thus obtained is integrated in the integrator 14 to determine the command position of each axis. And
This command position is output to the servo calculation unit 15, control calculation is performed in the servo calculation unit 15 so that the robot arm follows, and the calculated operation amount is output to the amplifier 16.
The motor 17 is driven, and the arm 19 is passed through the speed reducer 18.
Is driven. The rotation position of the motor 17 is the detector 2
It is detected at 0 and is fed back to the servo calculation unit 15.

The acceleration / deceleration processing is intended to suppress the occurrence of vibration of the mechanical system at the time of starting and stopping the robot by smoothing the speed change in the normalized speed pattern.

That is, for example, based on an acceleration pattern showing a trapezoidal acceleration change as shown in FIG. 2A, a speed pattern of a smooth speed change in which an acceleration peak is suppressed is obtained, and control is performed according to this speed pattern. This is to suppress the occurrence of vibration.

However, in reality, the vibration cannot be effectively reduced by such a method.

That is, it is considered that the commanded acceleration corresponds to the driving force acting on the arm. When the power spectrum distribution is obtained by Fourier transforming the acceleration in the acceleration pattern of FIG. 2A, the distribution is shown in FIG. It can be seen that as shown in (b), many components of the resonance frequency fn of the mechanical system are included. It is considered that this component acts as a vibrating force and causes the arm to vibrate when the robot is started and stopped.

Therefore, in this embodiment, instead of directly obtaining the velocity pattern from the trapezoidal acceleration pattern,
The frequency analysis of the acceleration pattern is performed as described below to obtain the normalized speed pattern.

That is, in the apparatus of the embodiment, a predetermined acceleration pattern, for example, a trapezoidal waveform acceleration pattern as shown in FIG. 2A is prepared, and the acceleration of this acceleration pattern is Fourier transformed to obtain a power spectrum distribution (see FIG. 2).
(B)) is obtained, and processing for removing frequency components near the mechanical system resonance frequency fn from this power spectrum distribution is performed. As a result, a power spectrum distribution having no frequency component near the mechanical resonance frequency fn as shown in FIG. 3A is obtained. As the part to be removed here, the resonance frequency of the robot varies depending on its operating position, that is, the possible posture, and the width of the frequency to be removed varies, so that the component of the varying width a may be removed. desirable.

Then, the power spectrum distribution shown in FIG. 3A is subjected to the inverse Fourier transform, and the acceleration pattern as shown in FIG. 3B is recreated. Then, the acceleration in this acceleration pattern is integrated to create a normalized velocity pattern as shown in FIG.

In the embodiment, since the acceleration / deceleration processing is performed by using the normalized speed pattern having no resonance frequency component as described above, the arm at the time of starting / stopping the robot is different from the acceleration / deceleration processing already adopted. The vibration of is greatly reduced.

Further, as described above, when the resonance frequency component is removed, the part corresponding to the possible posture of the robot is removed in advance, so it is not necessary to obtain the resonance frequency online every time the posture of the robot changes. Therefore, the problem of increase in the amount of calculation as described above is also solved.

Further, it has been known that the distribution of frequency components contained in the acceleration pattern is greatly affected by the acceleration / deceleration time. Therefore, as will be described later, a plurality of normalized speed patterns are obtained and stored in advance according to the magnitude of the acceleration / deceleration time, and thereafter, a plurality of normalized speed patterns corresponding to the calculated acceleration / deceleration times are stored. If only the data is read out from the inside, the processing will be performed immediately for any teaching speed.

Based on the above, the apparatus of the embodiment shown in FIG. 1 will be described in more detail with reference to FIGS. 4 and 5 again.

The normalized speed pattern from which the resonance frequency component has been removed as described above, as shown in FIG. 5, has patterns P1 and P1 corresponding to a plurality of acceleration / deceleration times t1 to tN.
2, P3 ... PN are stored in the normalized speed pattern memory 4.

On the other hand, the stored contents of the acceleration / deceleration time selection tables 2a and 2b are shown in FIGS. 4A and 4B, respectively, and are stored as the acceleration / deceleration times corresponding to the teaching speeds. In the case of PTP control, the acceleration / deceleration times tp1, tp2, ... Tp5 corresponding to the teaching speeds ω1, ω2, ... Acceleration / deceleration times tc1, tc2 ... tcN corresponding to teaching velocities v1, v2 ... VN indicating the moving speed of the tool tip of the robot are stored in the table 2b. Here, ωmax and vmax respectively indicate the maximum speed, and the acceleration / deceleration times tp1 to tp5 and tc1 to tc5 are
This corresponds to any of the acceleration / deceleration times t1 to tN shown in FIG. The method of dividing the acceleration / deceleration times t1, t2, ... Is arbitrary, and the number of divisions and the like can be arbitrarily set according to the memory capacity.

In the following, description will be made assuming PTP control acceleration processing.

First, when the teaching speed ωT is given to the acceleration / deceleration time selection processing section 1a, the acceleration / deceleration time selection table 2a is referred to, which corresponds to the given teaching speed ωT as shown in FIG. 4 (a). The acceleration time tp3 is selected. Next, in the scale conversion unit 3a, the speed pattern P3 corresponding to the acceleration time tp3 is calculated from the stored contents of the normalized speed pattern memory 4.
Is read out, and the normalized speed of the vertical axis of this speed pattern is multiplied by the teaching speed ωT to obtain the command speed ω (i). The command speed ω (i) is integrated by the integrator 5 to obtain the command position θ (i). And the command position θ (i)
The axes are driven and controlled by the servo calculator 8 so that the arm follows the robot, and the robot operates without vibration.

In the case of deceleration, control is performed in the same manner as the acceleration processing described above by using a pattern in which the normalized speed pattern shown in FIG. 5 is symmetrical with respect to the time axis.

The CP control is performed in the same manner as the PTP control described above.

That is, the acceleration / deceleration time corresponding to the teaching speed is read based on FIG. 4B, the normalized speed pattern corresponding to the read time is read from FIG. 5, and this read. Each axis of the robot is controlled according to the normalized speed pattern.

[0038]

As described above, according to the present invention, the speed pattern is obtained based on the acceleration pattern from which the resonance frequency component of the mechanical system is removed, and the acceleration / deceleration processing is performed according to this speed pattern. Vibrations that occur during acceleration and deceleration of the robot can be effectively suppressed, and computations can be performed in a short time for real-time control.

[Brief description of drawings]

FIG. 1 is a control block diagram showing a configuration of an embodiment of a vibration reducing apparatus for a robot according to the present invention.

FIG. 2 is a graph used to explain the acceleration / deceleration processing of the embodiment.

FIG. 3 is a graph used to explain the acceleration / deceleration processing of the embodiment.

FIG. 4 is a graph showing stored contents of an acceleration / deceleration time selection table shown in FIG.

FIG. 5 is a graph showing the stored contents of the normalized speed pattern memory shown in FIG.

FIG. 6 is a control block diagram showing a configuration of a conventional device used for understanding an embodiment.

[Explanation of symbols]

 1a, 1b Acceleration / deceleration time selection processing unit 2a, 2b Acceleration / deceleration time selection table 3a, 3b Scale conversion unit 4 Normalized speed pattern memory

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI technical display location G05D 19/02 D 8610-3H

Claims (3)

[Claims] 1. A teaching speed of a tool tip of a robot or a driving speed of each axis of a robot is taught, and an acceleration time or a deceleration time until reaching the taught speed is calculated,
The speed pattern showing the time change of the speed until reaching the teaching speed is obtained within the calculated acceleration time or deceleration time, and the moving speed of the tool tip or the driving speed of each axis of the robot is determined according to the obtained speed pattern. In the robot configured to drive and control each axis of the robot until the teaching speed is reached, an acceleration pattern showing the time change of the acceleration until reaching the teaching speed is prepared, and the acceleration shown in this acceleration pattern is Fourier transformed. Then, the power spectrum distribution is obtained, the part corresponding to the frequency that excites the vibration of the robot is removed from this power spectrum distribution, and the remaining power spectrum distribution that has been removed is inverse Fourier transformed to recalculate the acceleration pattern. Integrating the acceleration shown in the corrected acceleration pattern A vibration reduction device for a robot, which is configured to calculate the speed pattern. 2. A plurality of speed patterns are previously calculated and prepared for each size of the acceleration time or the deceleration time, and correspond to the calculated time when the acceleration time or the deceleration time is calculated. The vibration reduction device for a robot according to claim 1, wherein the speed pattern to be performed is selected from the prepared plurality of speed patterns. 3. A range of frequencies for exciting vibrations of the robot is obtained for all possible postures of the robot,
The vibration reduction device for a robot according to claim 1, wherein a portion corresponding to the obtained frequency range is removed from the power spectrum distribution.