JPS63298501A - Controller - Google Patents

Abstract

PURPOSE:To prevent the generation of a large error to all repetitions by temporarily operating a learning function in a repetition operation at a low speed. CONSTITUTION:A variable speed learning controller 21 for operating at a lower speed than a desired repetition speed while the number of the repetitions is small to a repetition operation command is provided. Namely, in a first repetition, a speed setting device 22 sets the low speed, a variable speed command value generating device 20 generates a command value 7 corresponding to this speed and a memory 6 writes and reads at the rate proportional to this speed. After a second or first speed is repeated, the speed setting device 22 raises a set value to the desired speed. A controlled system 3 operates at the desired speed, however, since the quantity of operation is added by the quantity of compensation 10 previously obtained at the low speed, the large error is not generated. At the time of shifting to the operation at the desired speed, a model of controlled system is used to calculate a generalizing force and add to the quantity of a compensation, thereby, the error can be further reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a control device that gradually reduces errors during repeated operations without causing large errors.

[Traditional technique f nete]

Figure 4 shows a system equipped with a conventional learning controller. This is a configuration diagram of the TI device. In the figure, l is a command value generator, 2 is a PD controller, 3 is a controlled object, and 4 is a learning controller.

5 is an arithmetic unit, 6 is a memory, and 7 is a command value, r(t).

8 is the output, y, (t), 9 is the error, e((t).

10 is compensation t ui (t), 11 is operation 1. i is the number of repetitions.

Next, the operation will be explained.

FIG. 5 is a diagram showing an example of the response of the system shown in FIG. 4. Command value 7. For r (L), the output 8.3' t-+(t) at the i-ith repetition is as shown in the figure, and the compensation amount IO at this time is ul-1(t). do. In the i-th repetition, the compensation amount ui(t) is changed to ut
The value shown by the second term of equation (1) is added to -+(t).

u H(t) −u H−, (t) + −r e
1−+(t)−(1)t et (t)−r (t,) −y4−+(t)
H...(2) By choosing an appropriate value of r, i
The output 8.7i (t) of the ith repetition becomes the command value 7. r (t), and as a result, by making it an i-function, a! (t) converges to 0.

In the ii-th repetition, the arithmetic unit 11 calculates the error 9
ei-1(t) is input, and the second term of equation (1) is calculated.The output is added to the compensation amount ut-+(t) and stored in the memory 6. This value is read out in the next i-th repetition and output as a compensation amount of 10° ut(t). By repeating the same operations one after another,
The error 9 gradually decreases.

[Problem that the invention seeks to solve]

Since the conventional control device is configured as described above, the error 9 can be compensated to approach zero by increasing the number of repetitions! 10 can be obtained from equation (1), but in the first iteration, the compensation amount 10 calculated from equation (1) is zero, so there is no effect of the learning controller, and normal PDIi
tll? Similar to the control using only the IL unit, there is a problem in that the error 9 increases when high-speed operation is performed.

Also, in the second and subsequent iterations, the compensation ft1
There was a problem in that the error 9 was large, especially in high-speed operation, until o converged so that the error 9 was sufficiently close to zero.

The present invention has been made to solve the above-mentioned problems of conventional devices, and aims to provide a control device that does not cause large errors in all repetitions even when high-speed operation commands are given. It is something.

[Means for solving problems]

In order to achieve the above object, a control device according to the present invention includes a variable speed learning controller that operates at a speed lower than a desired repetition speed while the number of repetitions is small in response to a repetitive operation command.

[Effect]

The variable speed learning controller is operated at a low speed that does not increase the error until the error becomes small through repeated repetitions, so that no large error occurs in all repetitions.

[Embodiments of the invention]

FIG. 1 is a configuration diagram of a control device that is an embodiment of the present invention.
In the figure, 2 to 11 are the same or equivalent parts or functions as the conventional device, 20 is a variable speed command value generator, 21 is a variable speed learning controller, 22 is a speed setting device, and 23 and 24 are speed setting values. .

When a robot manipulator with six degrees of freedom is to be controlled, the equation of motion of the controlled object is given by equation (3).

matrix, 'lan is the inertia of the actuator, mj is the link mass, g is the gravitational acceleration + Fj is the position of the center of mass W vector, q is the generalized coordinate, F, l are the operation Ill of the controlled object
and the generalization power of the actuator, F.

This is the generalization power of Hamasatsu.

The third term on the right side of equation (3) is a term proportional to the square of velocity or acceleration, and can be made sufficiently smaller than the other terms by operating at a lower speed.

Therefore, by performing learning control at a sufficiently low speed, values corresponding to the fourth and fifth terms of the right-hand variation of equation (3), which do not depend on the speed, can be obtained as a complement (N amount 10).

Next, when operating at the desired speed, the amount of compensation obtained in advance at low speed is compensated for by shortening the time axis! MR10. As a result, large errors unlike conventional learning controllers that operate at a desired speed from the beginning will not occur.

In the first repetition, the speed setter 22 (
3) Set a low speed so that the terms up to the third term on the second right side are sufficiently smaller than the other terms. It is assumed that the shift command value generator 20 generates a command value 7 corresponding to this speed, and that the memory 6 performs writing and reading at a rate proportional to this speed.

The second time, or after repeating the first speed for a while, the speed setter 22 increases the set value to the desired speed. Controlled object 3 operates at the desired speed, but operation 1
Since the compensation amount 10 obtained at low speed is added to tll in advance, large errors unlike conventional learning controllers will not occur.

In addition, when shifting to operation at a desired speed, using the model of the controlled object, calculate the generalized force corresponding to the terms up to the third term on the right side of equation (3) and add it to the compensation amount 10. According to
Furthermore, it is possible to operate with fewer errors than in the past.

FIG. 2 is an embodiment of a control device equipped with a variable speed learning device 21 including a dynamics calculator 25 that calculates (3) up to the third term on the second right side, and the transition to a desired speed is achieved by the speed setting value 26. is added to the output of the memory 6 and output as the compensation amount 10.

In addition, in the above embodiment, after operating at a low speed where the speed-dependent term on the right side of equation (3) is sufficiently small compared to other terms, the speed is immediately increased to the desired speed and the operation is performed at a desired speed. It is also possible to gradually increase the speed up to the maximum speed. At this time, the values of the terms up to the third term on the right side of equation (3) gradually change, but the previous compensation amount 10 and error 9 are multiplied by an appropriate coefficient corresponding to the change in speed to compensate for the change in speed. By setting the amount to 10, the speed can be gradually increased to the desired speed without causing a large error.

FIG. 3 shows an embodiment of a control device equipped with a variable speed learning device 21 including a multiplier 27. When increasing the speed, the output of the memory 6 is multiplied by an appropriate coefficient corresponding to the change in speed. Multiply by 27, compensation 15tl.

It is said that

[Effects of the Invention] As described above, the control device according to the present invention is configured such that the learning control in the repetitive operation is performed once at a low speed, so that the control device according to the present invention does not cause a large error for all repetitions. It became a scale that provided the control device.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a control device according to an embodiment of the present invention, FIG. 2 is a block diagram showing a control device according to another embodiment of the present invention, and FIG. 3 is a block diagram showing a control device according to another embodiment of the present invention. FIG. 4 is a block diagram of a certain control device, FIG. 4 is a block diagram of a conventional control device, and FIG. 5 is a line diagram showing an example of a response. In the figure, 2 is the PD controller, 3 is the controlled object, 5 is the computing unit, 6 is the memory, 7 is the command value, 8 is the output, 9 is the error, 10 is the compensation amount, 11 is the vibration amount, and 20 is the variable speed command 21 is a variable speed learning controller, 22 is a speed setter, 23, 24, 26 is a speed setting value, 25 is a dynamics calculator, and 27 is a multiplier. Note that the same reference numerals in each figure indicate the same or corresponding parts. Agent: Souji Sasaki, Patent Attorney 6, Contents of amendment Procedural amendment (voluntary) 1. Indication of the case Japanese Patent Application No. 1983-131576 2. Name control device for the invention 3. Person making the amendment Relationship with the case Patent application Address 2-2-3 Marunouchi, Chiyoda-ku, Tokyo Name (601) Mitsubishi Electric Corporation Representative Moriya Shiki 4 Address of agent 5-8-6-5 Toranomon, Minato-ku, Tokyo Subject to amendment “Detailed Description of the Invention” column of the specification and drawings j =
nk -1a qk a qnJ Yoi
Correct fl qn to the following formula. (2) Figures 1 and 2 of the drawings are corrected to look like corrected drawings.

Claims (4)

[Claims] (1) means for repeating the operation and correcting the amount of correction so as to bring the output closer to the command value based on the amount of compensation and error in the i-1th repetition or earlier, and storing this amount of compensation using a memory; In the i-th repetition, in the learning control or repetition control that uses this compensation amount to reduce the error, the variable speed setting means is provided and the compensation amount is adjusted by operating at a speed lower than the desired operation speed at least once. A control device comprising a variable speed learning controller configured to correct. (2) The variable speed learning controller operates at least once at a speed lower than the desired operating speed, and shortens the time axis by assuming that the compensation amount obtained as a result is obtained at the desired speed. 2. The control device according to claim 1, wherein the control device is configured to use this correction amount to bring the output closer to the command value when moving to a desired speed. (3) The variable speed learning controller uses the model of the controlled object to calculate the amount of operation required to bring the output close to the command value at the desired operating speed, and when shifting to the desired operating speed, 3. The control device according to claim 2, wherein the control device is configured to add the value to the compensation amount obtained at low speed. (4) When the variable speed learning controller increases the operating speed for the i-th time, it modifies the compensation amount based on the compensation amount and error at the i-1st time, further shortens the time axis, and adjusts the compensation amount for the i-th time. 2. The control device according to claim 1, wherein the control device is configured to: